fix(subagents): deterministic announce delivery with descendant gating#35080
Merged
fix(subagents): deterministic announce delivery with descendant gating#35080
Conversation
openclaw-barnacle
bot
added
agents
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 82b2dfbd06
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
src/agents/subagent-announce.ts
Outdated
Comment on lines
+1174
to
+1176
| if (typeof listSubagentRunsForRequester === "function") { | ||
| const directChildren = listSubagentRunsForRequester(params.childSessionKey); | ||
| if (Array.isArray(directChildren) && directChildren.length > 0) { |
There was a problem hiding this comment.
listSubagentRunsForRequester(params.childSessionKey) pulls all historical child runs for that session key, then this branch replaces the current run output with synthesized child findings. For session-mode subagents that handle multiple tasks over time, a later run with no new children will still re-announce stale child results from prior runs, and the current roundOneReply/latest reply is ignored. This regresses announce accuracy by leaking old completions into new task updates.
Useful? React with 👍 / 👎.
Contributor
Greptile SummaryThis PR overhauls the subagent announcement delivery system to make parent-facing announce events deterministic by gating them on descendant completion rather than using speculative, potentially-stale direct sends. Key architectural changes:
Two items worth a closer look:
Confidence Score: 4/5
Last reviewed commit: 82b2dfb |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
src/agents/subagent-announce.ts
Outdated
Comment on lines
+1163
to
+1172
| const { countPendingDescendantRuns, listSubagentRunsForRequester } = | ||
| await loadSubagentRegistryRuntime(); | ||
| pendingChildDescendantRuns = Math.max(0, countPendingDescendantRuns(params.childSessionKey)); | ||
| if (pendingChildDescendantRuns > 0) { | ||
| // Deterministic nested announce policy: if this run still has unfinished | ||
| // descendants, do not announce yet. Wait for descendant cleanup retries | ||
| // to re-trigger this announce check once everything is complete. | ||
| shouldDeleteChildSession = false; | ||
| return false; | ||
| } |
Contributor
There was a problem hiding this comment.
Descendant deferral runs before run-mode ignore check, causing unnecessary retry cycles
When the announcing run has pending descendants AND the requester is a run-mode session that has already ended, this guard returns false (defer) rather than immediately dropping the announce. The shouldIgnorePostCompletionAnnounceForSession check only runs later, after descendants settle. This wastes retry budget (up to MAX_ANNOUNCE_RETRY_COUNT or ANNOUNCE_COMPLETION_HARD_EXPIRY_MS) waiting for grandchildren before eventually discarding the announcement in the run-ended early-exit path.
Moving the ignore check to immediately before (or immediately after) countPendingDescendantRuns would fast-path out without deferring:
// Fast-path: if requester is a run-mode ended session, drop now.
if (requesterIsSubagent) {
try {
const { shouldIgnorePostCompletionAnnounceForSession } = await loadSubagentRegistryRuntime();
if (shouldIgnorePostCompletionAnnounceForSession(targetRequesterSessionKey)) {
return true;
}
} catch {}
}
const { countPendingDescendantRuns, listSubagentRunsForRequester } = await loadSubagentRegistryRuntime();
pendingChildDescendantRuns = Math.max(0, countPendingDescendantRuns(params.childSessionKey));
if (pendingChildDescendantRuns > 0) { ... }As currently written, the correctness is preserved (the announce is eventually dropped), but unnecessary retries accumulate when a run-mode requester has already ended and the announcing child has its own pending descendants.
Prompt To Fix With AI
This is a comment left during a code review.
Path: src/agents/subagent-announce.ts
Line: 1163-1172
Comment:
**Descendant deferral runs before run-mode ignore check, causing unnecessary retry cycles**
When the announcing run has pending descendants AND the requester is a run-mode session that has already ended, this guard returns `false` (defer) rather than immediately dropping the announce. The `shouldIgnorePostCompletionAnnounceForSession` check only runs later, after descendants settle. This wastes retry budget (up to `MAX_ANNOUNCE_RETRY_COUNT` or `ANNOUNCE_COMPLETION_HARD_EXPIRY_MS`) waiting for grandchildren before eventually discarding the announcement in the run-ended early-exit path.
Moving the ignore check to immediately before (or immediately after) `countPendingDescendantRuns` would fast-path out without deferring:
```typescript
// Fast-path: if requester is a run-mode ended session, drop now.
if (requesterIsSubagent) {
try {
const { shouldIgnorePostCompletionAnnounceForSession } = await loadSubagentRegistryRuntime();
if (shouldIgnorePostCompletionAnnounceForSession(targetRequesterSessionKey)) {
return true;
}
} catch {}
}
const { countPendingDescendantRuns, listSubagentRunsForRequester } = await loadSubagentRegistryRuntime();
pendingChildDescendantRuns = Math.max(0, countPendingDescendantRuns(params.childSessionKey));
if (pendingChildDescendantRuns > 0) { ... }
```
As currently written, the correctness is preserved (the announce is eventually dropped), but unnecessary retries accumulate when a run-mode requester has already ended and the announcing child has its own pending descendants.
How can I resolve this? If you propose a fix, please make it concise.
src/agents/subagent-announce.ts
Outdated
Comment on lines
+1174
to
+1188
| if (typeof listSubagentRunsForRequester === "function") { | ||
| const directChildren = listSubagentRunsForRequester(params.childSessionKey); | ||
| if (Array.isArray(directChildren) && directChildren.length > 0) { | ||
| childCompletionFindings = buildChildCompletionFindings( | ||
| directChildren.map((child) => ({ | ||
| childSessionKey: child.childSessionKey, | ||
| task: child.task, | ||
| label: child.label, | ||
| createdAt: child.createdAt, | ||
| endedAt: child.endedAt, | ||
| frozenResultText: child.frozenResultText, | ||
| outcome: child.outcome, | ||
| })), | ||
| ); | ||
| } |
Contributor
There was a problem hiding this comment.
listSubagentRunsForRequester may include multiple run records for session-mode children
listRunsForRequesterFromRuns returns ALL run records where requesterSessionKey === childSessionKey, including every historical run for session-mode children that have been reused across multiple tasks. If a session-mode child processed 3 sequential tasks during the parent's lifetime, all 3 run records would appear in buildChildCompletionFindings, making the announce message verbose and potentially confusing.
Consider deduplicating to only the LATEST completed run per childSessionKey when building child completion findings:
const latestPerChild = new Map<string, SubagentRunRecord>();
for (const child of directChildren) {
const existing = latestPerChild.get(child.childSessionKey);
if (!existing || child.createdAt > existing.createdAt) {
latestPerChild.set(child.childSessionKey, child);
}
}
childCompletionFindings = buildChildCompletionFindings([...latestPerChild.values()].map(...));This is only relevant for session-mode subagents that run multiple tasks; run-mode subagents use unique session keys per run and are unaffected.
Prompt To Fix With AI
This is a comment left during a code review.
Path: src/agents/subagent-announce.ts
Line: 1174-1188
Comment:
**`listSubagentRunsForRequester` may include multiple run records for session-mode children**
`listRunsForRequesterFromRuns` returns ALL run records where `requesterSessionKey === childSessionKey`, including every historical run for session-mode children that have been reused across multiple tasks. If a session-mode child processed 3 sequential tasks during the parent's lifetime, all 3 run records would appear in `buildChildCompletionFindings`, making the announce message verbose and potentially confusing.
Consider deduplicating to only the LATEST completed run per `childSessionKey` when building child completion findings:
```typescript
const latestPerChild = new Map<string, SubagentRunRecord>();
for (const child of directChildren) {
const existing = latestPerChild.get(child.childSessionKey);
if (!existing || child.createdAt > existing.createdAt) {
latestPerChild.set(child.childSessionKey, child);
}
}
childCompletionFindings = buildChildCompletionFindings([...latestPerChild.values()].map(...));
```
This is only relevant for session-mode subagents that run multiple tasks; run-mode subagents use unique session keys per run and are unaffected.
How can I resolve this? If you propose a fix, please make it concise.
tyler6204
added a commit
that referenced
this pull request
Mar 5, 2026
tyler6204
force-pushed
the
fix/subagent-announce-delivery
branch
from
0f5040a to
2002e47
Compare
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 2002e47a43
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| baselineReply: reply, | ||
| maxWaitMs: Math.max(minReplyChangeWaitMs, Math.min(params.timeoutMs, 2_000)), | ||
| }); | ||
| if (!childCompletionFindings) { |
There was a problem hiding this comment.
The isAnnounceSkip and isSilentReplyText guards are currently nested under if (!childCompletionFindings), so they are skipped whenever direct-child findings are synthesized. In nested/session orchestrator flows, this means a subagent reply of NO_REPLY (for duplicate or late completion events) can still be forwarded upstream as a fresh announce, producing duplicate user-facing updates instead of being suppressed. These suppression checks should run regardless of whether childCompletionFindings is present.
Useful? React with 👍 / 👎.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 7a08f3f77b
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Comment on lines
+331
to
+333
| if (entry.frozenResultText !== undefined) { | ||
| return false; | ||
| } |
There was a problem hiding this comment.
freezeRunResultAtCompletion permanently short-circuits once frozenResultText is set, but the first snapshot can legitimately miss output because completion can be observed before transcript writes settle (the announce flow itself documents this race) and captureSubagentCompletionReply only polls briefly (1.5s in src/agents/subagent-announce.ts). If that first capture stores null, later retries never refresh it, so parent synthesis (buildChildCompletionFindings) can report (no output) for children that actually produced a final reply.
Useful? React with 👍 / 👎.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 1cada50138
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
src/agents/subagent-announce.ts
Outdated
Comment on lines
+1252
to
+1254
| pendingChildDescendantRuns = Math.max(0, countPendingDescendantRuns(params.childSessionKey)); | ||
| if (pendingChildDescendantRuns > 0) { | ||
| // Deterministic nested announce policy: if this run still has unfinished |
There was a problem hiding this comment.
runSubagentAnnounceFlow now unconditionally blocks announce delivery when countPendingDescendantRuns(params.childSessionKey) > 0, but cron dispatch calls this flow with childSessionKey = params.agentSessionKey (see src/cron/isolated-agent/delivery-dispatch.ts), i.e. the main session rather than a subagent run. That means any unrelated in-flight subagent under the main session can suppress a cron completion announce and make cron delivery report failure even though the cron run itself finished, which is a user-visible delivery regression.
Useful? React with 👍 / 👎.
Comment on lines
+34
to
+38
| const requesterRun = requesterRunId ? runs.get(requesterRunId) : undefined; | ||
| const requesterRunMatchesScope = | ||
| requesterRun && requesterRun.childSessionKey === key ? requesterRun : undefined; | ||
| const lowerBound = requesterRunMatchesScope?.startedAt ?? requesterRunMatchesScope?.createdAt; | ||
| const upperBound = requesterRunMatchesScope?.endedAt; |
There was a problem hiding this comment.
When requesterRunId is missing from the registry (or does not match the requester session), this function silently drops both bounds and returns all historical child runs for that requester session; runSubagentAnnounceFlow then synthesizes findings from that list and can overwrite the real completion text. This is especially risky for cron announces because they pass a synthetic childRunId, so stale subagent history can leak into the cron payload instead of the cron output.
Useful? React with 👍 / 👎.
tyler6204
force-pushed
the
fix/subagent-announce-delivery
branch
from
1cada50 to
bb0b35a
Compare
🔒 Aisle Security AnalysisWe found 4 potential security issue(s) in this PR:
1. 🟠 Unscoped child-run listing when requesterRunId is missing can leak stale outputs into announce synthesis
DescriptionThe new scoping logic in This can cause stale child outputs from prior runs to be included in a later completion announcement:
Vulnerable code: const requesterRunId = options?.requesterRunId?.trim();
const requesterRun = requesterRunId ? runs.get(requesterRunId) : undefined;
const requesterRunMatchesScope =
requesterRun && requesterRun.childSessionKey === key ? requesterRun : undefined;
const lowerBound = requesterRunMatchesScope?.startedAt ?? requesterRunMatchesScope?.createdAt;
const upperBound = requesterRunMatchesScope?.endedAt;
return [...runs.values()].filter((entry) => {
if (entry.requesterSessionKey !== key) {
return false;
}
if (typeof lowerBound === "number" && entry.createdAt < lowerBound) {
return false;
}
if (typeof upperBound === "number" && entry.createdAt > upperBound) {
return false;
}
return true;
});When RecommendationFail closed when Option A (recommended): return an empty list if the requester run cannot be resolved/matched. export function listRunsForRequesterFromRuns(
runs: Map<string, SubagentRunRecord>,
requesterSessionKey: string,
options?: { requesterRunId?: string },
): SubagentRunRecord[] {
const key = requesterSessionKey.trim();
if (!key) return [];
const requesterRunId = options?.requesterRunId?.trim();
if (requesterRunId) {
const requesterRun = runs.get(requesterRunId);
if (!requesterRun || requesterRun.childSessionKey !== key) {
// Do not silently fall back to unscoped listing.
return [];
}
const lowerBound = requesterRun.startedAt ?? requesterRun.createdAt;
const upperBound = requesterRun.endedAt;
return [...runs.values()].filter((entry) =>
entry.requesterSessionKey === key &&
entry.createdAt >= lowerBound &&
(typeof upperBound !== "number" || entry.createdAt <= upperBound)
);
}
return [...runs.values()].filter((entry) => entry.requesterSessionKey === key);
}Option B: Change the API to pass explicit bounds (startedAt/endedAt) from the caller, rather than deriving them from a lookup that can fail. Additionally, consider storing an explicit 2. 🟡 Prompt injection via unescaped UNTRUSTED_CHILD_RESULT delimiters in internal event prompt text
Description
Vulnerable code: "<<<BEGIN_UNTRUSTED_CHILD_RESULT>>>",
event.result || "(no output)",
"<<<END_UNTRUSTED_CHILD_RESULT>>>",Impact depends on the downstream model’s instruction-following behavior, but this formatting makes delimiter-confusion attacks easier because the boundary markers are not escaped/encoded and are trivial to reproduce. RecommendationDo not embed untrusted text in a delimiter scheme that can be closed by the untrusted text itself. Mitigations (pick one, strongest first):
Example (escaping sentinel strings): function escapeUntrustedBlock(text: string): string {
return text
.replaceAll("<<<BEGIN_UNTRUSTED_CHILD_RESULT>>>", "<<<BEGIN_UNTRUSTED_CHILD_RESULT_ESCAPED>>>")
.replaceAll("<<<END_UNTRUSTED_CHILD_RESULT>>>", "<<<END_UNTRUSTED_CHILD_RESULT_ESCAPED>>>");
}
// usage
const safeResult = escapeUntrustedBlock(event.result || "(no output)");
lines.push(
"<<<BEGIN_UNTRUSTED_CHILD_RESULT>>>",
safeResult,
"<<<END_UNTRUSTED_CHILD_RESULT>>>",
);Also consider additionally placing untrusted output in a code-fenced block and prefixing each line (e.g., 3. 🟡 Cron interim-ack retry can re-run agent turn and duplicate side effects
Description
Impact:
Vulnerable code (retry trigger and second execution): if (shouldRetryInterimAck) {
const continuationPrompt = [
"Your previous response was only an acknowledgement and did not complete this cron task.",
"Complete the original task now.",
...
].join(" ");
await runPrompt(continuationPrompt);
}Why the current guardrails are insufficient:
RecommendationMake the “interim-ack retry” safe-by-default by preventing any second execution after potential side effects, or make it explicitly opt-in per job. Recommended mitigation options (prefer multiple):
const continuationPrompt = [
"Continue from where you left off in this same session.",
"Before using any tool, review what you already did in the previous turn.",
"Do NOT repeat any action that was already completed (avoid duplicate sends/writes/deletes).",
"If actions were already performed, only produce the final summary now.",
].join(" ");
These changes reduce the risk of duplicate mutations while still avoiding "on it" cron responses. 4. 🔵 Sensitive subagent completion outputs persisted to disk in subagent registry (runs.json)
Description
Why this is a security/privacy issue:
Vulnerable code (capture + persistence trigger): const captured = await captureSubagentCompletionReply(entry.childSessionKey);
entry.frozenResultText = captured?.trim() ? capFrozenResultText(captured) : null;
...
persistSubagentRuns();And if (role === "toolResult" || role === "tool") {
return extractToolResultText((message as ToolResultMessage).content);
}Registry persistence writes the whole saveJsonFile(pathname, out);RecommendationReduce exposure of sensitive data at rest by changing what is persisted and for how long:
Example (avoid persisting tool output + clear after announce): // 1) Capture assistant-only text (no toolResult), and
// 2) Clear after announce even when cleanup=keep
const captured = await captureAssistantOnlyCompletion(entry.childSessionKey);
entry.frozenResultText = captured?.trim() ? capFrozenResultText(captured) : null;
// ... after didAnnounce === true
entry.frozenResultText = undefined;
entry.frozenResultCapturedAt = undefined;
persistSubagentRuns();Analyzed PR: #35080 at commit Last updated on: 2026-03-06T04:10:53Z |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 5e922098b4
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
src/agents/subagent-registry.ts
Outdated
| if (shouldDeleteAttachments) { | ||
| await safeRemoveAttachmentsDir(entry); | ||
| } | ||
| clearFrozenRunResult(entry); |
There was a problem hiding this comment.
Clearing frozenResultText during successful child cleanup causes nested parent synthesis to lose data: runSubagentAnnounceFlow builds parent findings from listSubagentRunsForRequester(...).frozenResultText, but parent announce is deferred until descendants reach cleanupCompletedAt, which means this cleanup path has already run for those children. For run-mode subagents (default cleanup: "keep"), this turns real child outputs into (no output) in the parent’s final synthesized announce.
Useful? React with 👍 / 👎.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: f95fc20070
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Comment on lines
+34
to
+38
| const requesterRun = requesterRunId ? runs.get(requesterRunId) : undefined; | ||
| const requesterRunMatchesScope = | ||
| requesterRun && requesterRun.childSessionKey === key ? requesterRun : undefined; | ||
| const lowerBound = requesterRunMatchesScope?.startedAt ?? requesterRunMatchesScope?.createdAt; | ||
| const upperBound = requesterRunMatchesScope?.endedAt; |
There was a problem hiding this comment.
listRunsForRequesterFromRuns silently drops its time bounds when options.requesterRunId is absent or not found, so callers get every historical child run for that requester session. In this commit, runSubagentAnnounceFlow now trusts that scoped result for childCompletionFindings, so contexts like cron dispatch (which passes a synthetic childRunId) can replace the current completion text with stale historical child outputs and announce incorrect results.
Useful? React with 👍 / 👎.
| } | ||
| } | ||
|
|
||
| if (!childCompletionFindings) { |
There was a problem hiding this comment.
The NO_REPLY/silent suppression checks only run inside if (!childCompletionFindings), so they are skipped whenever descendant findings are synthesized. If a subagent run intentionally finished with NO_REPLY (for duplicate/late completion handling) but has child findings available, this path still emits a new announce upstream, causing duplicate user-visible updates instead of suppression.
Useful? React with 👍 / 👎.
Comment on lines
+361
to
+363
| if (entry.frozenResultText !== undefined) { | ||
| return false; | ||
| } |
There was a problem hiding this comment.
freezeRunResultAtCompletion permanently exits once frozenResultText is set, including the null case from an early empty capture. Because completion capture only polls briefly, a race where transcript writes land after that first attempt leaves frozenResultText stuck at null, and later parent synthesis reports (no output) even though the child produced a final reply.
Useful? React with 👍 / 👎.
tyler6204
force-pushed
the
fix/subagent-announce-delivery
branch
from
f5023f7 to
04a5383
Compare
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 04a5383e4b
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| } | ||
| if ( | ||
| (hadActiveDescendants || expectedSubagentFollowup) && | ||
| hadDescendants && |
There was a problem hiding this comment.
The stale-interim suppression now depends on hadDescendants, but that flag is only true when descendants are currently active or a fallback reply was recovered. If the first cron reply indicates subagent follow-up (expectedSubagentFollowup === true) and descendants already settled without a recoverable text payload (for example all descendants return NO_REPLY or their transcript is unavailable), hadDescendants stays false and the later guard will forward the original interim text ("on it", "spawned a subagent") to users instead of suppressing it. This is a regression from the previous behavior that treated expected follow-up itself as sufficient to suppress stale interim output.
Useful? React with 👍 / 👎.
tyler6204
force-pushed
the
fix/subagent-announce-delivery
branch
from
04a5383 to
bb412c6
Compare
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: bb412c61b1
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Comment on lines
+320
to
+322
| const immediate = await readLatestSubagentOutput(sessionKey); | ||
| if (immediate?.trim()) { | ||
| return immediate; |
There was a problem hiding this comment.
captureSubagentCompletionReply returns the first non-empty transcript entry immediately, but for session-mode subagents the latest assistant reply can still belong to the previous task when a new run finishes before transcript writes settle. Because freezeRunResultAtCompletion stores this value and runSubagentAnnounceFlow later trusts it as roundOneReply, the completion announce for the current run can report the previous run’s output. This shows up in reused subagent sessions handling sequential tasks quickly and causes incorrect user-facing completion text.
Useful? React with 👍 / 👎.
tyler6204
force-pushed
the
fix/subagent-announce-delivery
branch
from
bb412c6 to
0520506
Compare
tyler6204
force-pushed
the
fix/subagent-announce-delivery
branch
from
0520506 to
a5f639c
Compare
There was a problem hiding this comment.
💡 Codex Review
openclaw/src/agents/subagent-announce.ts
Lines 791 to 795 in a5f639c
Reintroduce direct send for finalized completion payloads
expectsCompletionMessage flows now always invoke agent (sendSubagentAnnounceDirectly) instead of using the old completion-direct send path, so cron/subagent completions that already have finalized text (roundOneReply) depend on a second model run in the target session. In contexts where that session cannot run a model turn (auth/profile mismatch, model outage, or transient run failure), delivery fails even though the final payload was already computed; this directly contradicts the cron caller’s expectation in src/cron/isolated-agent/delivery-dispatch.ts that completion mode performs direct outbound delivery.
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Comment on lines
+756
to
+759
| const effectiveDirectOrigin = | ||
| params.expectsCompletionMessage && completionDirectOrigin | ||
| ? completionDirectOrigin | ||
| : directOrigin; |
There was a problem hiding this comment.
For completion announcements, effectiveDirectOrigin always prefers completionDirectOrigin when present, even if it lacks a deliverable destination; then shouldDeliverExternally becomes false and the call succeeds as an internal agent turn without sending externally. This is a regression from the prior logic that still used directOrigin for the agent path when completion-specific targeting was unusable, and it can silently suppress user delivery when a hook/bound override returns a partial origin.
Useful? React with 👍 / 👎.
This was referenced Mar 6, 2026
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 39a2081771
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| } else { | ||
| didAnnounce = delivery.delivered; | ||
| } | ||
| didAnnounce = delivery.delivered; |
There was a problem hiding this comment.
Setting didAnnounce directly from delivery.delivered treats queue/steer fallback as a successful cron delivery, because runSubagentAnnounceDispatch marks queued outcomes as delivered. In the cron path (dispatchCronDelivery), that causes runs to be recorded as delivered even when the direct outbound send failed and only an internal queue/injection happened, which can silently mask real channel delivery failures in non-best-effort jobs.
Useful? React with 👍 / 👎.
Comment on lines
+408
to
+412
| for (const entry of candidates) { | ||
| if (entry.frozenResultText === nextFrozen) { | ||
| continue; | ||
| } | ||
| entry.frozenResultText = nextFrozen; |
There was a problem hiding this comment.
This loop applies one captured session reply to every pending completion record for that session, so concurrent pending runs in a reused subagent session can overwrite each other’s frozen outputs. When a later untracked lifecycle end arrives (for example from another turn in the same session), earlier pending runs can be re-labeled with the newer reply, causing incorrect child-result synthesis and wrong announce text for those older runs.
Useful? React with 👍 / 👎.
This was referenced Mar 6, 2026
|
Deterministic subagent announce delivery with descendant gating is crucial for preventing race conditions and ensuring proper agent coordination. This addresses some real safety concerns in multi-agent systems: Delivery ordering guarantees:
Failure handling:
Security implications:
Observability:
Resource management:
This kind of coordination primitive is essential for building reliable agent systems at scale. Many teams reinvent this poorly. |
thinstripe
pushed a commit
to thinstripe/openclaw
that referenced
this pull request
Mar 6, 2026
…lt refresh, and cron retry (openclaw#35080) Thanks @tyler6204
handsdiff
pushed a commit
to handsdiff/activeclaw
that referenced
this pull request
Mar 6, 2026
…lt refresh, and cron retry (openclaw#35080) Thanks @tyler6204
jenawant
pushed a commit
to jenawant/openclaw
that referenced
this pull request
Mar 10, 2026
…lt refresh, and cron retry (openclaw#35080) Thanks @tyler6204
V-Gutierrez
pushed a commit
to V-Gutierrez/openclaw-vendor
that referenced
this pull request
Mar 17, 2026
…lt refresh, and cron retry (openclaw#35080) Thanks @tyler6204
alexey-pelykh
pushed a commit
to remoteclaw/remoteclaw
that referenced
this pull request
Mar 20, 2026
…lt refresh, and cron retry (openclaw#35080) Thanks @tyler6204 (cherry picked from commit 81b93b9)
alexey-pelykh
pushed a commit
to remoteclaw/remoteclaw
that referenced
this pull request
Mar 20, 2026
…lt refresh, and cron retry (openclaw#35080) Thanks @tyler6204 (cherry picked from commit 81b93b9)
9 tasks
0x666c6f
added a commit
to 0x666c6f/openclaw
that referenced
this pull request
Mar 26, 2026
…nclaw#29) * Plugins/llm-task: migrate to scoped plugin-sdk imports * Plugins/lobster: migrate to scoped plugin-sdk imports * Plugins/matrix: migrate to scoped plugin-sdk imports * Plugins/mattermost: migrate to scoped plugin-sdk imports * Plugins/memory-core: migrate to scoped plugin-sdk imports * Plugins/memory-lancedb: migrate to scoped plugin-sdk imports * Plugins/minimax-portal-auth: migrate to scoped plugin-sdk imports * Plugins/msteams: migrate to scoped plugin-sdk imports * Plugins/nextcloud-talk: migrate to scoped plugin-sdk imports * Plugins/nostr: migrate to scoped plugin-sdk imports * Plugins/open-prose: migrate to scoped plugin-sdk imports * Plugins/phone-control: migrate to scoped plugin-sdk imports * Plugins/qwen-portal-auth: migrate to scoped plugin-sdk imports * Plugins/synology-chat: migrate to scoped plugin-sdk imports * Plugins/talk-voice: migrate to scoped plugin-sdk imports * Plugins/test-utils: migrate to scoped plugin-sdk imports * Plugins/thread-ownership: migrate to scoped plugin-sdk imports * Plugins/tlon: migrate to scoped plugin-sdk imports * Plugins/twitch: migrate to scoped plugin-sdk imports * Plugins/voice-call: migrate to scoped plugin-sdk imports * Plugins/whatsapp: migrate to scoped plugin-sdk imports * Plugins/zalo: migrate to scoped plugin-sdk imports * Plugins/zalouser: migrate to scoped plugin-sdk imports * Chore: remove accidental .DS_Store artifact * chore(docs): add plugins refactor changelog entry * feat(ios): add Live Activity connection status + stale cleanup (#33591) * feat(ios): add live activity connection status and cleanup Add lock-screen/Dynamic Island connection health states and prune duplicate/stale activities before reuse. This intentionally excludes AI/title generation and heavier UX rewrites from #27488. Co-authored-by: leepokai <[email protected]> * fix(ios): treat ended live activities as inactive * chore(changelog): add PR reference and author thanks --------- Co-authored-by: leepokai <[email protected]> * fix: kill stuck ACP child processes on startup and harden sessions in discord threads (#33699) * Gateway: resolve agent.wait for chat.send runs * Discord: harden ACP thread binding + listener timeout * ACPX: handle already-exited child wait * Gateway/Discord: address PR review findings * Discord: keep ACP error-state thread bindings on startup * gateway: make agent.wait dedupe bridge event-driven * discord: harden ACP probe classification and cap startup fan-out * discord: add cooperative timeout cancellation * discord: fix startup probe concurrency helper typing * plugin-sdk: avoid Windows root-alias shard timeout * plugin-sdk: keep root alias reflection path non-blocking * discord+gateway: resolve remaining PR review findings * gateway+discord: fix codex review regressions * Discord/Gateway: address Codex review findings * Gateway: keep agent.wait lifecycle active with shared run IDs * Discord: clean up status reactions on aborted runs * fix: add changelog note for ACP/Discord startup hardening (#33699) (thanks @dutifulbob) --------- Co-authored-by: Onur <[email protected]> * fix: relay ACP sessions_spawn parent streaming (#34310) (thanks @vincentkoc) (#34310) Co-authored-by: Onur Solmaz <[email protected]> * fix(telegram): materialize dm draft final to avoid duplicates * docs(changelog): credit @Brotherinlaw-13 for #34318 * fix: prevent nodes media base64 context bloat (#34332) * fix: preserve raw media invoke for HTTP tool clients (#34365) * fix(slack): route system events to bound agent sessions (#34045) * fix(slack): route system events via binding-aware session keys * fix(slack): pass sender to system event session resolver * fix(slack): include sender context for interaction session routing * fix(slack): include modal submitter in session routing * test(slack): cover binding-aware system event routing * test(slack): update interaction session key assertions * test(slack): assert reaction session routing carries sender * docs(changelog): note slack system event routing fix * Update CHANGELOG.md * Delete changelog/fragments directory * fix(memory): serialize local embedding initialization to avoid duplicate model loads (#15639) Merged via squash. Prepared head SHA: a085fc21a8ba7163fffdb5de640dd4dc1ff5a88e Co-authored-by: SubtleSpark <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(model): propagate custom provider headers to model objects (#27490) Merged via squash. Prepared head SHA: e4183b398fc7eb4c18b2b691cb0dd882ec993608 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(daemon): handle systemctl is-enabled exit 4 (not-found) on Ubuntu (#33634) Merged via squash. Prepared head SHA: 67dffc3ee239cd7b813cb200c3dd5475d9e203a6 Co-authored-by: Yuandiaodiaodiao <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(node-host): sync rawCommand with hardened argv after executable path pinning (#33137) Merged via squash. Prepared head SHA: a7987905f7ad6cf5fee286ffa81ceaad8297174f Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * Agents: add generic poll-vote action support * fix(ollama): pass provider headers to Ollama stream function (#24285) createOllamaStreamFn() only accepted baseUrl, ignoring custom headers configured in models.providers.<provider>.headers. This caused 403 errors when Ollama endpoints are behind reverse proxies that require auth headers (e.g. X-OLLAMA-KEY via HAProxy). Add optional defaultHeaders parameter to createOllamaStreamFn() and merge them into every fetch request. Provider headers from config are now passed through at the call site in the embedded runner. Fixes #24285 * test(ollama): add default header precedence coverage * chore(changelog): add PR entry openclaw#24337 thanks @echoVic * Outbound: allow text-only plugin adapters * Outbound: avoid empty multi-media fallback sends * chore(changelog): align outbound adapter entry openclaw#32788 thanks @liuxiaopai-ai * fix(outbound): fail media-only text-only adapter fallback * chore(changelog): clarify outbound media-only fallback openclaw#32788 thanks @liuxiaopai-ai * fix(review): enforce behavioral sweep validation * Fix gateway restart false timeouts on Debian/systemd (#34874) * daemon(systemd): target sudo caller user scope * test(systemd): cover sudo user scope commands * infra(ports): fall back to ss when lsof missing * test(ports): verify ss fallback listener detection * cli(gateway): use probe fallback for restart health * test(gateway): cover restart-health probe fallback * Compaction/Safeguard: require structured summary headings (#25555) Merged via squash. Prepared head SHA: 0b1df34806a7b788261290be55760fd89220de53 Co-authored-by: rodrigouroz <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * Fix Linux daemon install checks when systemd user bus env is missing (#34884) * daemon(systemd): fall back to machine user scope when user bus is missing * test(systemd): cover machine scope fallback for user-bus errors * test(systemd): reset execFile mock state across cases * test(systemd): make machine-user fallback assertion portable * fix(daemon): keep root sudo path on direct user scope * test(systemd): cover sudo root user-scope behavior * ci: use resolvable bun version in setup-node-env * agents: preserve totalTokens on request failure instead of using contextWindow (#34275) Merged via squash. Prepared head SHA: f9d111d0a79a07815d476356e98a28df3a0000ba Co-authored-by: RealKai42 <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix: align AGENTS.md template section names with post-compaction extraction (#25029) (#25098) Merged via squash. Prepared head SHA: 8cd6cc8049aab5a94d8a9d5fb08f2e792c4ac5fd Co-authored-by: echoVic <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * Changelog: add daemon systemd user-bus fallback entry (#34884) * Changelog: add gateway restart health entry (#34874) * fix: finalize spanish locale support * fix: add spanish locale support (#35038) (thanks @DaoPromociones) * fix(deps): patch hono transitive audit vulnerabilities * fix(security): avoid prototype-chain account path checks (#34982) Merged via squash. Prepared head SHA: f89cc6a649959997fe1dec1e1c1bff9a61b2de98 Co-authored-by: HOYALIM <[email protected]> Co-authored-by: dvrshil <[email protected]> Reviewed-by: @dvrshil * fix(deps): bump tar to 7.5.10 * docs(changelog): document dependency security fixes * fix: restore auto-reply system events timeline (#34794) (thanks @anisoptera) (#34794) Co-authored-by: Ayaan Zaidi <[email protected]> * fix(feishu): comprehensive reply mechanism — outbound replyToId forwarding + topic-aware reply targeting (#33789) * fix(feishu): comprehensive reply mechanism fix — outbound replyToId forwarding + topic-aware reply targeting - Forward replyToId from ChannelOutboundContext through sendText/sendMedia to sendMessageFeishu/sendMarkdownCardFeishu/sendMediaFeishu, enabling reply-to-message via the message tool. - Fix group reply targeting: use ctx.messageId (triggering message) in normal groups to prevent silent topic thread creation (#32980). Preserve ctx.rootId targeting for topic-mode groups (group_topic/group_topic_sender) and groups with explicit replyInThread config. - Add regression tests for both fixes. Fixes #32980 Fixes #32958 Related #19784 * fix: normalize Feishu delivery.to before comparing with messaging tool targets - Add normalizeDeliveryTarget helper to strip user:/chat: prefixes for Feishu - Apply normalization in matchesMessagingToolDeliveryTarget before comparison - This ensures cron duplicate suppression works when session uses prefixed targets (user:ou_xxx) but messaging tool extract uses normalized bare IDs (ou_xxx) Fixes review comment on PR #32755 (cherry picked from commit fc20106f16ccc88a5f02e58922bb7b7999fe9dcd) * fix(feishu): catch thrown SDK errors for withdrawn reply targets The Feishu Lark SDK can throw exceptions (SDK errors with .code or AxiosErrors with .response.data.code) for withdrawn/deleted reply targets, in addition to returning error codes in the response object. Wrap reply calls in sendMessageFeishu and sendCardFeishu with try-catch to handle thrown withdrawn/not-found errors (230011, 231003) and fall back to client.im.message.create, matching the existing response-level fallback behavior. Also extract sendFallbackDirect helper to deduplicate the direct-send fallback block across both functions. Closes #33496 (cherry picked from commit ad0901aec103a2c52f186686cfaf5f8ba54b4a48) * feishu: forward outbound reply target context (cherry picked from commit c129a691fcf552a1cebe1e8a22ea8611ffc3b377) * feishu extension: tighten reply target fallback semantics (cherry picked from commit f85ec610f267020b66713c09e648ec004b2e26f1) * fix(feishu): align synthesized fallback typing and changelog attribution * test(feishu): cover group_topic_sender reply targeting --------- Co-authored-by: Xu Zimo <[email protected]> Co-authored-by: Munem Hashmi <[email protected]> Co-authored-by: bmendonca3 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(feishu): use msg_type media for mp4 video (fixes #33674) (#33720) * fix(feishu): use msg_type media for mp4 video (fixes #33674) * Feishu: harden streaming merge semantics and final reply dedupe Use explicit streaming update semantics in the Feishu reply dispatcher: treat onPartialReply payloads as snapshot updates and block fallback payloads as delta chunks, then merge final text with the shared overlap-aware mergeStreamingText helper before closing the stream. Prevent duplicate final text delivery within the same dispatch cycle, and add regression tests covering overlap snapshot merge, duplicate final suppression, and block-as-delta behavior to guard against repeated/truncated output. * fix(feishu): prefer message.reply for streaming cards in topic threads * fix: reduce Feishu streaming card print_step to avoid duplicate rendering Fixes openclaw/openclaw#33751 * Feishu: preserve media sends on duplicate finals and add media synthesis changelog * Feishu: only dedupe exact duplicate final replies * Feishu: use scoped plugin-sdk import in streaming-card tests --------- Co-authored-by: 倪汉杰0668001185 <[email protected]> Co-authored-by: zhengquanliu <[email protected]> Co-authored-by: nick <[email protected]> Co-authored-by: linhey <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(agents): bypass pendingDescendantRuns guard for cron announce delivery (#35185) * fix(agents): bypass pendingDescendantRuns guard for cron announce delivery Standalone cron job completions were blocked from direct channel delivery when the cron run had spawned subagents that were still registered as pending. The pendingDescendantRuns guard exists for live orchestration coordination and should not apply to fire-and-forget cron announce sends. Thread the announceType through the delivery chain and skip both the child-descendant and requester-descendant pending-run guards when the announce originates from a cron job. Closes #34966 * fix: ensure outbound session entry for cron announce with named agents (#32432) Named agents may not have a session entry for their delivery target, causing the announce flow to silently fail (delivered=false, no error). Two fixes: 1. Call ensureOutboundSessionEntry when resolving the cron announce session key so downstream delivery can find channel metadata. 2. Fall back to direct outbound delivery when announce delivery fails to ensure cron output reaches the target channel. Closes #32432 Co-Authored-By: Claude Opus 4.6 <[email protected]> * fix: guard announce direct-delivery fallback against suppression leaks (#32432) The `!delivered` fallback condition was too broad — it caught intentional suppressions (active subagents, interim messages, SILENT_REPLY_TOKEN) in addition to actual announce delivery failures. Add an `announceDeliveryWasAttempted` flag so the direct-delivery fallback only fires when `runSubagentAnnounceFlow` was actually called and failed. Also remove the redundant `if (route)` guard in `resolveCronAnnounceSessionKey` since `resolved` being truthy guarantees `route` is non-null. Co-Authored-By: Claude Opus 4.6 <[email protected]> * fix(cron): harden announce synthesis follow-ups --------- Co-authored-by: scoootscooob <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * Feishu: harden streaming merge semantics and final reply dedupe (#33245) * Feishu: close duplicate final gap and cover routing precedence * Feishu: resolve reviewer duplicate-final and routing feedback * Feishu: tighten streaming send-mode option typing * Feishu: fix reverse-overlap streaming merge ordering * Feishu: align streaming final dedupe test expectation * Feishu: allow distinct streaming finals while deduping repeats --------- Co-authored-by: Tak Hoffman <[email protected]> * fix: cron backup should preserve pre-edit snapshot (#35195) (#35234) * fix(cron): avoid overwriting .bak during normalization Fixes openclaw/openclaw#35195 * test(cron): preserve pre-edit bak snapshot in normalization path --------- Co-authored-by: 0xsline <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(cron): stabilize restart catch-up replay semantics (#35351) * Cron: stabilize restart catch-up replay semantics * Cron: respect backoff in startup missed-run replay * cron: narrow startup replay backoff guard (#35391) * cron: unify stale-run recovery and preserve manual-run every anchors (#35363) * cron: unify stale-run recovery and preserve manual every anchors * cron: address unresolved review threads on recovery paths * cron: remove duplicate timestamp helper after rebase * refactor(telegram): remove unused webhook callback helper (#27816) * fix(pr): make review claim step required * fix(skills): deduplicate slash commands by skillName across all interfaces Move skill-command deduplication by skillName from the Discord-only `dedupeSkillCommandsForDiscord` into `listSkillCommandsForAgents` so every interface (TUI, Slack, text) consistently sees a clean command list without platform-specific workarounds. When multiple agents share a skill with the same name the old code emitted `github` + `github_2` and relied on Discord to collapse them. Now `listSkillCommandsForAgents` returns only the first registration per skillName, and the Discord-specific wrapper is removed. Co-Authored-By: Claude Sonnet 4.6 <[email protected]> * style: fix formatting in skill-commands.test.ts and provider.ts Co-Authored-By: Claude Sonnet 4.6 <[email protected]> * style(skills): align formatting cleanup for dedupe changes * chore(changelog): add dedupe note openclaw#27521 thanks @shivama205 * fix(agents): detect Venice provider proxying xAI/Grok models for schema cleaning (#35355) Merged via squash. Prepared head SHA: 8bfdec257bb6a6025cb69a0a213a433da32b15db Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(agents): decode HTML entities in xAI/Grok tool call arguments (#35276) Merged via squash. Prepared head SHA: c4445d2938898ded9c046614f9315dbda65ec573 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(agents): guard promoteThinkingTagsToBlocks against malformed content entries (#35143) Merged via squash. Prepared head SHA: 3971122f5fd27c66c8c9c5ce783f00e113b1f47b Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(web-ui): render Accounts schema node properly (#35380) Co-authored-by: stakeswky <[email protected]> Co-authored-by: liuxiaopai-ai <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(agents): guard context pruning against malformed thinking blocks (#35146) Merged via squash. Prepared head SHA: a196a565b1b8e806ffbf85172bcf1128796b45a2 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(gateway): prevent internal route leakage in chat.send Synthesis of routing fixes from #35321, #34635, and #35356 for internal-client reply safety. - Require explicit `deliver: true` before inheriting any external delivery route. - Keep webchat/TUI/UI-origin traffic on internal routing by default. - Allow configured-main session inheritance only for non-Webchat/UI clients, and honor `session.mainKey`. - Add regression tests for UI no-inherit, configured-main CLI inherit, and deliver-flag behavior. Co-authored-by: alexyyyander <[email protected]> Co-authored-by: Octane0411 <[email protected]> Co-authored-by: Linux2010 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(gateway): pass actual version to Control UI client instead of dev (#35230) * fix(gateway): pass actual version to Control UI client instead of "dev" The GatewayClient, CLI WS client, and browser Control UI all sent "dev" as their clientVersion during handshake, making it impossible to distinguish builds in gateway logs and health snapshots. - GatewayClient and CLI WS client now use the resolved VERSION constant - Control UI reads serverVersion from the bootstrap endpoint and forwards it when connecting - Bootstrap contract extended with serverVersion field Closes #35209 * Gateway: fix control-ui version version-reporting consistency * Control UI: guard deferred bootstrap connect after disconnect * fix(ui): accept same-origin http and relative gateway URLs for client version --------- Co-authored-by: Tak Hoffman <[email protected]> * chore(pr): enforce changelog placement and reduce merge sync churn * TTS: add baseUrl support to OpenAI TTS config (#34321) Merged via squash. Prepared head SHA: e9a10cf81d2021cf81091dfa81e13ffdbb6a540a Co-authored-by: RealKai42 <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * ACP: add persistent Discord channel and Telegram topic bindings (#34873) * docs: add ACP persistent binding experiment plan * docs: align ACP persistent binding spec to channel-local config * docs: scope Telegram ACP bindings to forum topics only * docs: lock bound /new and /reset behavior to in-place ACP reset * ACP: add persistent discord/telegram conversation bindings * ACP: fix persistent binding reuse and discord thread parent context * docs: document channel-specific persistent ACP bindings * ACP: split persistent bindings and share conversation id helpers * ACP: defer configured binding init until preflight passes * ACP: fix discord thread parent fallback and explicit disable inheritance * ACP: keep bound /new and /reset in-place * ACP: honor configured bindings in native command flows * ACP: avoid configured fallback after runtime bind failure * docs: refine ACP bindings experiment config examples * acp: cut over to typed top-level persistent bindings * ACP bindings: harden reset recovery and native command auth * Docs: add ACP bound command auth proposal * Tests: normalize i18n registry zh-CN assertion encoding * ACP bindings: address review findings for reset and fallback routing * ACP reset: gate hooks on success and preserve /new arguments * ACP bindings: fix auth and binding-priority review findings * Telegram ACP: gate ensure on auth and accepted messages * ACP bindings: fix session-key precedence and unavailable handling * ACP reset/native commands: honor fallback targets and abort on bootstrap failure * Config schema: validate ACP binding channel and Telegram topic IDs * Discord ACP: apply configured DM bindings to native commands * ACP reset tails: dispatch through ACP after command handling * ACP tails/native reset auth: fix target dispatch and restore full auth * ACP reset detection: fallback to active ACP keys for DM contexts * Tests: type runTurn mock input in ACP dispatch test * ACP: dedup binding route bootstrap and reset target resolution * reply: align ACP reset hooks with bound session key * docs: replace personal discord ids with placeholders * fix: add changelog entry for ACP persistent bindings (#34873) (thanks @dutifulbob) --------- Co-authored-by: Onur <[email protected]> * docs(telegram): recommend allowlist for single-user DM policy (#34841) * docs(telegram): recommend allowlist for single-user bots * docs(telegram): condense single-user allowlist note --------- Co-authored-by: echoVic <[email protected]> * fix(feishu): check response.ok before calling response.json() in streaming card (#35628) Merged via squash. Prepared head SHA: 62c3fec80d97cea9be344c0bef5358a0a5dc5560 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * Mattermost: honor onmessage mention override and add gating diagnostics tests (#27160) Merged via squash. Prepared head SHA: 6cefb1d5bf3d6dfcec36c1cee3f9ea887f10c890 Co-authored-by: turian <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Reviewed-by: @mukhtharcm * fix(subagents): strip leaked [[reply_to]] tags from completion announces (#34503) * fix(subagents): strip reply tags from completion delivery text * test(subagents): cover reply-tag stripping in cron completion sends * changelog: note iMessage reply-tag stripping in completion announces * Update CHANGELOG.md * Apply suggestion from @greptile-apps[bot] Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> --------- Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix(cron): restore direct fallback after announce failure in best-effort mode (openclaw#36177) Verified: - pnpm build - pnpm check (fails on pre-existing origin/main lint debt in extensions/mattermost imports) - pnpm test:macmini Co-authored-by: Tak Hoffman <[email protected]> * test(cron): add cross-channel announce fallback regression coverage (openclaw#36197) Verified: - pnpm install --frozen-lockfile - pnpm build - pnpm check (fails on pre-existing origin/main lint debt in extensions/mattermost imports) - pnpm test:macmini Co-authored-by: Tak Hoffman <[email protected]> * feat(mattermost): add interactive buttons support (#19957) Merged via squash. Prepared head SHA: 8a25e608729d0b9fd07bb0ee4219d199d9796dbe Co-authored-by: tonydehnke <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Reviewed-by: @mukhtharcm * fix(browser): remove deprecated --disable-blink-features=AutomationControlled flag - Removes OpenClaw's default `--disable-blink-features=AutomationControlled` Chrome launch switch to avoid unsupported-flag warnings in newer Chrome (#35721). - Preserves compatibility for older Chrome via `browser.extraArgs` override behavior (source analysis: #35770, #35728, #35727, #35885). - Synthesis attribution: thanks @Sid-Qin, @kevinWangSheng, @ningding97, @Naylenv, @clawbie. Source PR refs: #35734, #35770, #35728, #35727, #35885 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: kevinWangSheng <[email protected]> Co-authored-by: ningding97 <[email protected]> Co-authored-by: Naylenv <[email protected]> Co-authored-by: clawbie <[email protected]> Co-authored-by: Takhoffman <[email protected]> * fix(feishu): add HTTP timeout to prevent per-chat queue deadlocks (#36430) When the Feishu API hangs or responds slowly, the sendChain never settles, causing the per-chat queue to remain in a processing state forever and blocking all subsequent messages in that thread. This adds a 30-second default timeout to all Feishu HTTP requests by providing a timeout-aware httpInstance to the Lark SDK client. Closes #36412 Co-authored-by: Ayane <[email protected]> * fix(feishu): use probed botName for mention checks (#36391) * Feishu: honor bot mentions by ID despite aliases (Fixes #36317) (#36333) * Mattermost: switch plugin-sdk imports to scoped subpaths (openclaw#36480) Verified: - pnpm build - pnpm check - pnpm test:macmini Co-authored-by: Takhoffman <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(feishu): accept groupPolicy "allowall" as alias for "open" (#36358) * fix(feishu): accept groupPolicy "allowall" as alias for "open" When users configure groupPolicy: "allowall" in Feishu channel config, the Zod schema rejects the value and the runtime policy check falls through to the allowlist path. With an empty allowFrom array, all group messages are silently dropped despite the intended "allow all" semantics. Accept "allowall" at the schema level (transform to "open") and add a runtime guard in isFeishuGroupAllowed so the value is handled even if it bypasses schema validation. Closes #36312 Made-with: Cursor * Feishu: tighten allowall alias handling and coverage --------- Co-authored-by: Tak Hoffman <[email protected]> * synthesis: fix Feishu group mention slash parsing ## Summary\n\nFeishu group slash command parsing is fixed for mentions and command probes across authorization paths.\n\nThis includes:\n- Normalizing bot mention text in group context for reliable slash detection in message parsing.\n- Adding command-probe normalization for group slash invocations.\n\nCo-authored-by: Sid Qin <[email protected]>\nCo-authored-by: Tak Hoffman <[email protected]> * Feishu: normalize group slash command probing - Feishu/group slash command detection: normalize group mention wrappers before command-authorization probing so mention-prefixed commands are recognized in group routing.\n- Source PR: #36011\n- Contributor: @liuxiaopai-ai\n\nCo-authored-by: Tak Hoffman <[email protected]>\nCo-authored-by: liuxiaopai-ai <[email protected]> * add prependSystemContext and appendSystemContext to before_prompt_build (fixes #35131) (#35177) Merged via squash. Prepared head SHA: d9a2869ad69db9449336a2e2846bd9de0e647ac6 Co-authored-by: maweibin <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(feishu): avoid media regressions from global HTTP timeout (#36500) * fix(feishu): avoid media regressions from global http timeout * fix(feishu): source HTTP timeout from config * fix(feishu): apply media timeout override to image uploads * fix(feishu): invalidate cached client when timeout changes * fix(feishu): clamp timeout values and cover image download * Gateway: add SecretRef support for gateway.auth.token with auth-mode guardrails (#35094) * fix(embedded): classify model_context_window_exceeded as context overflow, trigger compaction (#35934) Merged via squash. Prepared head SHA: 20fa77289c80b2807a6779a3df70440242bc18ca Co-authored-by: RealKai42 <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(agents): skip compaction API call when session has no real messages (#36451) Merged via squash. Prepared head SHA: 52dd6317895c7bd10855d2bd7dbbfc2f5279b68e Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(ui): catch marked.js parse errors to prevent Control UI crash (#36445) - Prevent Control UI session render crashes when `marked.parse()` encounters pathological recursive markdown by safely falling back to escaped `<pre>` output. - Tighten markdown fallback regression coverage and keep changelog attribution in sync for this crash-hardening path. Co-authored-by: Bin Deng <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(session): archive old transcript on daily/scheduled reset to prevent orphaned files (#35493) Merged via squash. Prepared head SHA: 0d95549d752adecfc0b08d5cd55a8b8c75e264fe Co-authored-by: byungsker <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(agents): set preserveSignatures to isAnthropic in resolveTranscriptPolicy (#32813) Merged via squash. Prepared head SHA: f522d21ca59a42abac554435a0aa646f6a34698d Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix: avoid false global rate-limit classification from generic cooldown text (#32972) Merged via squash. Prepared head SHA: 813c16f5afce415da130a917d9ce9f968912b477 Co-authored-by: stakeswky <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * refactor(agents): share failover HTTP status classification (#36615) * fix(agents): classify transient failover statuses consistently * fix(agents): preserve legacy failover status mapping * fix(failover): narrow service-unavailable to require overload indicator (#32828) (#36646) Merged via squash. Prepared head SHA: 46fb4306127972d7635f371fd9029fbb9baff236 Co-authored-by: jnMetaCode <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * Compaction/Safeguard: add summary quality audit retries (#25556) Merged via squash. Prepared head SHA: be473efd1635616ebbae6e649d542ed50b4a827f Co-authored-by: rodrigouroz <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * test(agents): add provider-backed failover regressions (#36735) * test(agents): add provider-backed failover fixtures * test(agents): cover more provider error docs * test(agents): tighten provider doc fixtures * Docs: add Slack typing reaction fallback * Docs: update gateway config reference for Slack and TTS * Docs: clarify OpenAI-compatible TTS endpoints * Docs: document Control UI locale support * Docs: cover heartbeat, cron, and plugin route updates * fix(ui): bump dompurify to 3.3.2 (#36781) * UI: bump dompurify to 3.3.2 * Deps: refresh dompurify lockfile * UI: hoist lifecycle connect test mocks (#36788) * fix(agents): classify insufficient_quota 400s as billing (#36783) * feat: append UTC time alongside local time in shared Current time lines (#32423) Merged via squash. Prepared head SHA: 9e8ec13933b5317e7cff3f0bc048de515826c31a Co-authored-by: jriff <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * fix(auth): grant senderIsOwner for internal channels with operator.admin scope (openclaw#35704) Verified: - pnpm install --frozen-lockfile - pnpm build - pnpm check - pnpm test:macmini Co-authored-by: Naylenv <[email protected]> Co-authored-by: Octane0411 <[email protected]> Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(config): prevent RangeError in merged schema cache key generation Fix merged schema cache key generation for high-cardinality plugin/channel metadata by hashing incrementally instead of serializing one large aggregate string. Includes changelog entry for the user-visible regression fix. Co-authored-by: Tak Hoffman <[email protected]> Co-authored-by: Bill <[email protected]> * fix(slack): propagate mediaLocalRoots through Slack send path Restore Slack local file upload parity with CVE-era local media allowlist enforcement by threading `mediaLocalRoots` through the Slack send call chain. - pass `ctx.mediaLocalRoots` from Slack channel action adapter into `handleSlackAction` - add and forward `mediaLocalRoots` in Slack action context/send path - pass `mediaLocalRoots` into `sendMessageSlack` for upload allowlist enforcement - add changelog entry with attribution for this behavior fix Co-authored-by: 2233admin <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(slack): preserve dedupe while recovering dropped app_mention (#34937) This PR fixes Slack mention loss without reintroducing duplicate dispatches. - Preserve seen-message dedupe at ingress to prevent duplicate processing. - Allow a one-time app_mention retry only when the paired message event was previously dropped before dispatch. - Add targeted race tests for both recovery and duplicate-prevention paths. Co-authored-by: littleben <[email protected]> Co-authored-by: OpenClaw Agent <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * README: add algal to contributors list (#2046) * fix: decouple Discord inbound worker timeout from listener timeout (#36602) (thanks @dutifulbob) (#36602) Co-authored-by: Onur Solmaz <[email protected]> * plugins: enforce prompt hook policy with runtime validation (#36567) Merged via squash. Prepared head SHA: 6b9d883b6ae33628235fb02ce39c0d0f46a065bb Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(memory): avoid destructive qmd collection rebinds * Harden Telegram poll gating and schema consistency (#36547) Merged via squash. Prepared head SHA: f77824419e3d166f727474a9953a063a2b4547f2 Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(browser): close tracked tabs on session cleanup (#36666) * Diffs: restore system prompt guidance (#36904) Merged via squash. Prepared head SHA: 1b3be3c87957c068473d5c86b9efba4a1a8503f2 Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(routing): avoid full binding rescans in resolveAgentRoute (#36915) * fix(gateway): honor insecure ws override for remote hostnames * fix(llm-task): load runEmbeddedPiAgent from dist/extensionAPI in installs * fix(auth): harden openai-codex oauth login path * feat(telegram/acp): Topic Binding, Pin Binding Message, Fix Spawn Param Parsing (#36683) * fix(acp): normalize unicode flags and Telegram topic binding * feat(telegram/acp): restore topic-bound ACP and session bindings * fix(acpx): clarify permission-denied guidance * feat(telegram/acp): pin spawn bind notice in topics * docs(telegram): document ACP topic thread binding behavior * refactor(reply): share Telegram conversation-id resolver * fix(telegram/acp): preserve bound session routing semantics * fix(telegram): respect binding persistence and expiry reporting * refactor(telegram): simplify binding lifecycle persistence * fix(telegram): bind acp spawns in direct messages * fix: document telegram ACP topic binding changelog (#36683) (thanks @huntharo) --------- Co-authored-by: Onur <[email protected]> * fix(gateway): preserve streamed prefixes across tool boundaries * fix(tui): prevent stale model indicator after /model * Memory: handle SecretRef keys in doctor embeddings (#36835) Merged via squash. Prepared head SHA: c1a3d0caae60115d886e8bfc9983c9533c773f04 Co-authored-by: joshavant <[email protected]> Co-authored-by: joshavant <[email protected]> Reviewed-by: @joshavant * fix(openai-codex): request required oauth api scopes (#24720) * fix(memory-flush): ban timestamped variant files in default flush prompt (#34951) Merged via squash. Prepared head SHA: efadda4988b460e6da07be72994d4951d64239d0 Co-authored-by: zerone0x <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(tui): render final event error when assistant output is empty (#14687) * feat(agents): flush reply pipeline before compaction wait (#35489) Merged via squash. Prepared head SHA: 7dbbcc510b74b0e8d35eb750d24575e34b5d769a Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(secrets): harden api key normalization for ByteString headers * fix(slack): remove double mrkdwn conversion in native streaming path Remove redundant text normalization from Slack native streaming markdown_text flow so Markdown formatting is preserved. Synthesis context: overlaps reviewed from #34931, #34759, #34716, #34682, #34814. Co-authored-by: littleben <[email protected]> Co-authored-by: dunamismax <[email protected]> Co-authored-by: Octane <[email protected]> Co-authored-by: Mitsuyuki Osabe <[email protected]> Co-authored-by: Kai <[email protected]> Co-authored-by: OpenClaw Agent <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(kimi-coding): normalize anthropic tool payload format * fix(slack): thread channel ID through inbound context for reactions (#34831) Slack reaction/thread context routing fixes via canonical synthesis of #34831. Co-authored-by: Tak <[email protected]> * fix(heartbeat): pin HEARTBEAT.md reads to workspace path * fix(subagents): recover announce cleanup after kill/complete race * feat(hooks): emit compaction lifecycle hooks (#16788) * fix(auth): harden openai-codex oauth refresh fallback * fix(subagents): announce delivery with descendant gating, frozen result refresh, and cron retry (#35080) Thanks @tyler6204 * fix(agents): avoid synthetic tool-result writes on idle-timeout cleanup * fix(agent): harden undici stream timeouts for long openai-completions runs * fix(slack): record app_mention retry key before dedupe check (#37033) - Prime app_mention retry allowance before dedupe so near-simultaneous message/app_mention races do not drop valid mentions. - Prevent duplicate dispatch when app_mention wins the race and message prepare later succeeds. - Prune dispatched mention keys and add regression coverage for both dropped and successful in-flight message outcomes. Co-authored-by: Tak Hoffman <[email protected]> * fix(agents): honor explicit rate-limit cooldown probes in fallback runs * fix(agents): allow configured ollama endpoints without dummy api keys * fix(memory): recover qmd updates from duplicate document constraints * Doctor: warn on implicit heartbeat directPolicy (#36789) * Changelog: note heartbeat directPolicy doctor warning * Tests: cover heartbeat directPolicy doctor warning * Doctor: warn on implicit heartbeat directPolicy * Tests: cover per-agent heartbeat directPolicy warning * Update CHANGELOG.md * Plugins: clarify registerHttpHandler migration errors (#36794) * Changelog: note plugin HTTP route migration diagnostics * Tests: cover registerHttpHandler migration diagnostics * Plugins: clarify registerHttpHandler migration errors * Tests: cover registerHttpHandler diagnostic edge cases * Plugins: tighten registerHttpHandler migration hint * fix(memory): repair qmd collection name conflicts during ensure * fix(memory): handle qmd search results without docid * Plugins: avoid false integrity drift prompts on unpinned updates (#37179) * Plugins: skip drift prompts for unpinned updates * Plugins: cover unpinned integrity update behavior * Changelog: add #37179 release note * Delete changelog/fragments directory * Update CHANGELOG.md * fix(whatsapp): remove implicit [openclaw] self-chat prefix * fix: remove config.schema from agent gateway tool (#7382) Merged via squash. Prepared head SHA: f34a7780690a941936b31899e2d096b8a07f4afc Co-authored-by: kakuteki <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * feat(openai): add gpt-5.4 support for API and Codex OAuth (#36590) * feat(openai): add gpt-5.4 support and priority processing * feat(openai-codex): add gpt-5.4 oauth support * fix(openai): preserve provider overrides in gpt-5.4 fallback * fix(openai-codex): keep xhigh for gpt-5.4 default * fix(models): preserve configured overrides in list output * fix(models): close gpt-5.4 integration gaps * fix(openai): scope service tier to public api * fix(openai): complete prep followups for gpt-5.4 support (#36590) (thanks @dorukardahan) --------- Co-authored-by: Tyler Yust <[email protected]> * fix(tui): preserve credential-like tokens in render sanitization * CLI: make read-only SecretRef status flows degrade safely (#37023) * CLI: add read-only SecretRef inspection * CLI: fix read-only SecretRef status regressions * CLI: preserve read-only SecretRef status fallbacks * Docs: document read-only channel inspection hook * CLI: preserve audit coverage for read-only SecretRefs * CLI: fix read-only status account selection * CLI: fix targeted gateway fallback analysis * CLI: fix Slack HTTP read-only inspection * CLI: align audit credential status checks * CLI: restore Telegram read-only fallback semantics * chore(changelog): update for #37023 Signed-off-by: joshavant <[email protected]> * fix(agents): disable usage streaming chunks on non-native openai-completions * feat(nano-banana-pro): add --aspect-ratio flag to generate_image.py (#28159) * feat(nano-banana-pro): add --aspect-ratio flag to generate_image.py * Nano Banana: allow all supported aspect ratios * Docs: expand nano banana aspect ratio options --------- Co-authored-by: Vincent Koc <[email protected]> * fix(gateway): support image_url in OpenAI chat completions (#34068) * fix(gateway): parse image_url in openai chat completions * test(gateway): cover openai chat completions image_url flows * docs(changelog): note openai image_url chat completions fix (#17685) * fix(gateway): harden openai image_url parsing and limits * test(gateway): add openai image_url regression coverage * docs(changelog): expand #17685 openai chat completions note * Gateway: make OpenAI image_url URL fetch opt-in and configurable * Diagnostics: redact image base64 payload data in trace logs * Changelog: note OpenAI image_url hardening follow-ups * Gateway: enforce OpenAI image_url total budget incrementally * Gateway: scope OpenAI image_url extraction to the active turn * Update CHANGELOG.md * fix(agents): avoid xAI web_search tool-name collisions * fix: clear Telegram DM draft after materialize (#36746) (thanks @joelnishanth) * Fix Control UI duplicate iMessage replies for internal webchat turns (#36151) * Auto-reply: avoid routing external replies from internal webchat turns * Auto-reply tests: cover internal webchat non-routing with external origin metadata * Changelog: add Control UI iMessage duplicate-reply fix note * Auto-reply context: track explicit deliver routes * Gateway chat: mark explicit external deliver routes in context * Auto-reply: preserve explicit deliver routes for internal webchat turns * Auto-reply tests: cover explicit deliver routes from internal webchat turns * Gateway chat tests: assert explicit deliver route context tagging * fix: enforce 600 perms for cron store and run logs (#36078) * fix: enforce secure permissions for cron store and run logs * fix(cron): enforce dir perms and gate posix tests on windows * Cron store tests: cover existing directory permission hardening * Cron run-log tests: cover existing directory permission hardening * Changelog: note cron file permission hardening --------- Co-authored-by: linhey <[email protected]> Co-authored-by: Vincent Koc <[email protected]> * fix(tui): accept canonical session-key aliases in chat event routing * Gateway: normalize OpenAI stream chunk text * Gateway: coerce chat deliverable route boolean * fix(web_search): align brave language codes with API * Respect source channel for agent event surfacing (#36030) * fix(session): prefer webchat routes for direct ui turns (#37135) * Gateway: discriminate input sources * Cron: migrate legacy provider delivery hints * Cron: stabilize runs-one-shot migration tests * fix(memory): retry mcporter after Windows EINVAL spawn * fix(onboarding): guard daemon status probe on headless linux * Gateway: add path-scoped config schema lookup (#37266) Merged via squash. Prepared head SHA: 0c4d187f6fb66f2799d4047585d6368e433c883a Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * docs(changelog): add pr entry * fix(ci): restore protocol and schema checks (#37470) * Fix failover for zhipuai 1310 Weekly/Monthly Limit Exhausted (#33813) Merged via squash. Prepared head SHA: 3dc441e58de48913720cf7b6137fa761758d8344 Co-authored-by: zhouhe-xydt <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * fix(openai-codex-oauth): stop mutating authorize url scopes * Update CHANGELOG.md * fix(auth): remove bogus codex oauth responses probe * docs(changelog): fold codex oauth fix notes * docs(changelog): add codex oauth pr reference (#37558) * fix(security:PLA-697): block local secret scratch files from commits * feat: add dev EKS deployment for openclaw SRE container * fix: stabilize slack socket mode in dev eks runtime * fix(eks): reuse monitoring incident auth for openclaw-sre * feat(deploy): add grafana env guards and prod deploy wrapper * fix(security): redact tool outputs and enforce secret-safe runtime defaults * feat(sre): harden heartbeat routing and enrich triage signals * docs: add SRE hybrid intelligence design Three-layer architecture to improve bot reasoning quality: - Layer 1: Service knowledge (auto-discovery + overlays + incident memory) - Layer 2: Multi-stage reasoning chain (triage → hypothesize → causal chain → action plan → cross-review) - Layer 3: Incident learning loop (structured cards, overlay suggestions, feedback signals) Co-Authored-By: Claude Opus 4.6 <[email protected]> * feat(sandbox): add boundary cli to common runtime image * docs(sre): finalize hybrid intelligence design v19 after 18 Codex review rounds Iteratively hardened the design through 18 adversarial Codex review rounds, resolving 60+ findings (15+ CRITICAL, 45+ HIGH). Key fixes include: - Decoupled incident_id (pre-Step11 immutable) from card_id (LLM-derived) - Unified evidence/memory sanitization via shared _strip_instruction_tokens - Fixed rerun interval default (3600s > heartbeat 1800s) to prevent every-cycle re-runs - Added Phase 2 cross-review gap note and dual-column depth table - Overlay suggestion idempotency via deterministic suggestion_key + upsert - Decoupled chain timeout from legacy budget check - Added predicate alignment docs for L3 pre-check vs L3 gate - Adopted dynamic evidence_completeness denominator matching existing code - Added scope note clarifying design vs current implementation Co-Authored-By: Claude Opus 4.6 <[email protected]> * feat(sre): complete hybrid intelligence rollout and bot hardening * fix(sre): harden slack reasoning, dual chain, and deploy auth/runtime * ci: add ecr release pipeline and infra-helm image update trigger * ci: switch workflows to github-hosted runners * ci: allow manual ecr env target and use updater v4 * ci: remove windows jobs from workflow * chore(pla-678): update fallback auth fixture and plan docs * fix(pla-678): preserve control-ui host-header fallback in dev config * refactor(pla-678): remove local sre chart and require infra-helm * refactor(pla-678): remove direct eks deploy scripts * ci(pla-678): remove labeler workflow * ci(pla-678): refresh checks after disabling labeler * build: install boundary CLI in SRE runtime image * fix(slack:PLA-678): keep one thread lane across users * fix(sre-skill): enforce live db-query runbook in dev seed * fix(slack): ignore bot thread history when seeding new sessions * fix(slack): force retry on repeated thread questions * fix(sre-skill): require retry on repeated db requests * docs(sre-skill): make retry policy generic across repeated asks * fix(sre:PLA-678): enforce live linear ticket updates * fix(sre:PLA-678): use [PLATFORM] Backlog linear project * fix(sre:PLA-678): label linked linear tickets as openclaw-sre * fix(sre:PLA-678): auto-label linked linear tickets for auto-pr * test(sre:PLA-678): make tracking label test executable * fix(slack:PLA-678): remove status-final completion banner (#16) * fix(slack:PLA-678): enable direct file attachments from agent replies (#17) * fix(slack:PLA-678): remove status-final completion banner * fix(slack): enable direct file attachments from agent replies * refactor(sre:PLA-678): remove local deploy config mirror (#18) * fix: compact oversized auto-generated pr bodies * fix(sre): add built-in linear incident memory provider * feat(sre): add erpc helper with flo secret query * fix(sre): enforce canonical rpc.morpho.dev erpc endpoint * refactor(sre): remove local deploy config mirror * fix(slack:PLA-678): force DM replies into per-message threads (#19) * fix: compact oversized auto-generated pr bodies * fix(sre): add built-in linear incident memory provider * feat(sre): add erpc helper with flo secret query * fix(sre): enforce canonical rpc.morpho.dev erpc endpoint * refactor(sre): remove local deploy config mirror * fix(slack): force threaded replies for direct messages * feat(cron:PLA-678): use conversation history in self-improve runs (#20) * fix(slack:PLA-678): improve progress update readability (#21) * fix(ci:PLA-678): restore main release pipeline health (#22) * fix(ci:PLA-678): restore main release pipeline health * fix(ci:PLA-678): harden auto-response token fallback * fix(ci:PLA-678): pin setup-bun to published release * test(ci:PLA-678): align slack thread session expectations * fix(sre:PLA-678): harden Slack reply delivery and image release flow (#23) * fix(slack:PLA-678): harden thread-scoped reply delivery * fix(slack:PLA-678): satisfy lint after rebase * ci(release:PLA-678): move image release to cached ecr flow * ci(release:PLA-678): use official docker build actions * fix(slack:PLA-722): preserve progress updates during final-answer gating (#24) * fix(sre:PLA-724): restore ECR runtime image toolchain (#25) * fix(failover): classify HTTP 402 as rate_limit when payload indicates usage limit (#30484) (#36802) * fix(failover): classify HTTP 402 as rate_limit when payload indicates usage limit (#30484) Some providers (notably Anthropic Claude Max plan) surface temporary usage/rate-limit failures as HTTP 402 instead of 429. Before this change, all 402s were unconditionally mapped to 'billing', which produced a misleading 'run out of credits' warning for Max plan users who simply hit their usage window. This follows the same pattern introduced for HTTP 400 in #36783: check the error message for an explicit rate-limit signal before falling back to the default status-code classification. - classifyFailoverReasonFromHttpStatus now returns 'rate_limit' for 402 when isRateLimitErrorMessage matches the payload text - Added regression tests covering both the rate-limit and billing paths on 402 * fix: narrow 402 rate-limit matcher to prevent billing misclassification The original implementation used isRateLimitErrorMessage(), which matches phrases like 'quota exceeded' that legitimately appear in billing errors. This commit replaces it with a narrow, 402-specific matcher that requires BOTH retry language (try again/retry/temporary/cooldown) AND limit terminology (usage limit/rate limit/organization usage). Prevents misclassification of errors like: 'HTTP 402: exceeded quota, please add credits' -> billing (not rate_limit) Added regression test for the ambiguous case. --------- Co-authored-by: Val Alexander <[email protected]> * fix(mattermost): allow reachable interaction callback URLs (#37543) Merged via squash. Prepared head SHA: 4d593731be5a5dcbf3106d596b38acfeb8cf0aa8 Co-authored-by: mukhtharcm <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Reviewed-by: @mukhtharcm --------- Signed-off-by: joshavant <[email protected]> Co-authored-by: Gustavo Madeira Santana <[email protected]> Co-authored-by: Mariano <[email protected]> Co-authored-by: leepokai <[email protected]> Co-authored-by: Bob <[email protected]> Co-authored-by: Onur <[email protected]> Co-authored-by: Ayaan Zaidi <[email protected]> Co-authored-by: Vincent Koc <[email protected]> Co-authored-by: huangcj <[email protected]> Co-authored-by: gumadeiras <[email protected]> Co-authored-by: Sid <[email protected]> Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Co-authored-by: a <[email protected]> Co-authored-by: echoVic <[email protected]> Co-authored-by: Shakker <[email protected]> Co-authored-by: liuxiaopai-ai <[email protected]> Co-authored-by: Rodrigo Uroz <[email protected]> Co-authored-by: rodrigouroz <[email protected]> Co-authored-by: jalehman <[email protected]> Co-authored-by: Kai <[email protected]> Co-authored-by: RealKai42 <[email protected]> Co-authored-by: 青雲 <[email protected]> Co-authored-by: echoVic <[email protected]> Co-authored-by: Darshil <[email protected]> Co-authored-by: Ho Lim <[email protected]> Co-authored-by: dvrshil <[email protected]> Co-authored-by: Isis Anisoptera <[email protected]> Co-authored-by: Madoka <[email protected]> Co-authored-by: Xu Zimo <[email protected]> Co-authored-by: Munem Hashmi <[email protected]> Co-authored-by: bmendonca3 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> Co-authored-by: Nhj <[email protected]> Co-authored-by: 倪汉杰0668001185 <[email protected]> Co-authored-by: zhengquanliu <[email protected]> Co-authored-by: nick <[email protected]> Co-authored-by: linhey <[email protected]> Co-authored-by: scoootscooob <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> Co-authored-by: rexl2018 <[email protected]> Co-authored-by: sline <[email protected]> Co-authored-by: 0xsline <[email protected]> Co-authored-by: Harold Hunt <[email protected]> Co-authored-by: Shivam <[email protected]> Co-authored-by: 不做了睡大觉 <[email protected]> Co-authored-by: alexyyyander <[email protected]> Co-authored-by: alexyyyander <[email protected]> Co-authored-by: Octane0411 <[email protected]> Co-authored-by: Linux2010 <[email protected]> Co-authored-by: echoVic <[email protected]> Co-authored-by: Joseph Turian <[email protected]> Co-authored-by: turian <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> Co-authored-by: Tony Dehnke <[email protected]> Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: kevinWangSheng <[email protected]> Co-authored-by: ningding97 <[email protected]> Co-authored-by: Naylenv <[email protected]> Co-authored-by: clawbie <[email protected]> Co-authored-by: Takhoffman <[email protected]> Co-authored-by: Ayane <[email protected]> Co-authored-by: Ayane <[email protected]> Co-authored-by: StingNing <[email protected]> Co-authored-by: maweibin <[email protected]> Co-authored-by: maweibin <[email protected]> Co-authored-by: Josh Avant <[email protected]> Co-authored-by: Bin Deng <[email protected]> Co-authored-by: Byungsker <[email protected]> Co-authored-by: altaywtf <[email protected]> Co-authored-by: Altay <[email protected]> Co-authored-by: jiangnan <[email protected]> Co-authored-by: jnMetaCode <[email protected]> Co-authored-by: Jacob Riff <[email protected]> Co-authored-by: jriff <[email protected]> Co-authored-by: Naylenv <[email protected]> Co-authored-by: Bill <[email protected]> Co-authored-by: 2233admin <[email protected]> Co-authored-by: 2233admin <[email protected]> Co-authored-by: littleben <[email protected]> Co-authored-by: littleben <[email protected]> Co-authored-by: OpenClaw Agent <[email protected]> Co-authored-by: Vignesh Natarajan <[email protected]> Co-authored-by: zerone0x <[email protected]> Co-authored-by: zerone0x <[email protected]> Co-authored-by: dunamismax <[email protected]> Co-authored-by: Octane <[email protected]> Co-authored-by: Mitsuyuki Osabe <[email protected]> Co-authored-by: Tak <[email protected]> Co-authored-by: Tyler Yust <[email protected]> Co-authored-by: Hinata Kaga (samon) <[email protected]> Co-authored-by: dorukardahan <[email protected]> Co-authored-by: Tyler Yust <[email protected]> Co-authored-by: Brenner Spear <[email protected]> Co-authored-by: aerelune <[email protected]> Co-authored-by: Frank Yang <[email protected]> Co-authored-by: zhouhe-xydt <[email protected]> Co-authored-by: zhouhe-xydt <[email protected]> Co-authored-by: Vignesh <[email protected]> Co-authored-by: OpenClaw SRE Bot <[email protected]> Co-authored-by: Xinhua Gu <[email protected]> Co-authored-by: Val Alexander <[email protected]>
0x666c6f
added a commit
to 0x666c6f/openclaw
that referenced
this pull request
Mar 26, 2026
) * Plugins/matrix: migrate to scoped plugin-sdk imports * Plugins/mattermost: migrate to scoped plugin-sdk imports * Plugins/memory-core: migrate to scoped plugin-sdk imports * Plugins/memory-lancedb: migrate to scoped plugin-sdk imports * Plugins/minimax-portal-auth: migrate to scoped plugin-sdk imports * Plugins/msteams: migrate to scoped plugin-sdk imports * Plugins/nextcloud-talk: migrate to scoped plugin-sdk imports * Plugins/nostr: migrate to scoped plugin-sdk imports * Plugins/open-prose: migrate to scoped plugin-sdk imports * Plugins/phone-control: migrate to scoped plugin-sdk imports * Plugins/qwen-portal-auth: migrate to scoped plugin-sdk imports * Plugins/synology-chat: migrate to scoped plugin-sdk imports * Plugins/talk-voice: migrate to scoped plugin-sdk imports * Plugins/test-utils: migrate to scoped plugin-sdk imports * Plugins/thread-ownership: migrate to scoped plugin-sdk imports * Plugins/tlon: migrate to scoped plugin-sdk imports * Plugins/twitch: migrate to scoped plugin-sdk imports * Plugins/voice-call: migrate to scoped plugin-sdk imports * Plugins/whatsapp: migrate to scoped plugin-sdk imports * Plugins/zalo: migrate to scoped plugin-sdk imports * Plugins/zalouser: migrate to scoped plugin-sdk imports * Chore: remove accidental .DS_Store artifact * chore(docs): add plugins refactor changelog entry * feat(ios): add Live Activity connection status + stale cleanup (#33591) * feat(ios): add live activity connection status and cleanup Add lock-screen/Dynamic Island connection health states and prune duplicate/stale activities before reuse. This intentionally excludes AI/title generation and heavier UX rewrites from #27488. Co-authored-by: leepokai <[email protected]> * fix(ios): treat ended live activities as inactive * chore(changelog): add PR reference and author thanks --------- Co-authored-by: leepokai <[email protected]> * fix: kill stuck ACP child processes on startup and harden sessions in discord threads (#33699) * Gateway: resolve agent.wait for chat.send runs * Discord: harden ACP thread binding + listener timeout * ACPX: handle already-exited child wait * Gateway/Discord: address PR review findings * Discord: keep ACP error-state thread bindings on startup * gateway: make agent.wait dedupe bridge event-driven * discord: harden ACP probe classification and cap startup fan-out * discord: add cooperative timeout cancellation * discord: fix startup probe concurrency helper typing * plugin-sdk: avoid Windows root-alias shard timeout * plugin-sdk: keep root alias reflection path non-blocking * discord+gateway: resolve remaining PR review findings * gateway+discord: fix codex review regressions * Discord/Gateway: address Codex review findings * Gateway: keep agent.wait lifecycle active with shared run IDs * Discord: clean up status reactions on aborted runs * fix: add changelog note for ACP/Discord startup hardening (#33699) (thanks @dutifulbob) --------- Co-authored-by: Onur <[email protected]> * fix: relay ACP sessions_spawn parent streaming (#34310) (thanks @vincentkoc) (#34310) Co-authored-by: Onur Solmaz <[email protected]> * fix(telegram): materialize dm draft final to avoid duplicates * docs(changelog): credit @Brotherinlaw-13 for #34318 * fix: prevent nodes media base64 context bloat (#34332) * fix: preserve raw media invoke for HTTP tool clients (#34365) * fix(slack): route system events to bound agent sessions (#34045) * fix(slack): route system events via binding-aware session keys * fix(slack): pass sender to system event session resolver * fix(slack): include sender context for interaction session routing * fix(slack): include modal submitter in session routing * test(slack): cover binding-aware system event routing * test(slack): update interaction session key assertions * test(slack): assert reaction session routing carries sender * docs(changelog): note slack system event routing fix * Update CHANGELOG.md * Delete changelog/fragments directory * fix(memory): serialize local embedding initialization to avoid duplicate model loads (#15639) Merged via squash. Prepared head SHA: a085fc21a8ba7163fffdb5de640dd4dc1ff5a88e Co-authored-by: SubtleSpark <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(model): propagate custom provider headers to model objects (#27490) Merged via squash. Prepared head SHA: e4183b398fc7eb4c18b2b691cb0dd882ec993608 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(daemon): handle systemctl is-enabled exit 4 (not-found) on Ubuntu (#33634) Merged via squash. Prepared head SHA: 67dffc3ee239cd7b813cb200c3dd5475d9e203a6 Co-authored-by: Yuandiaodiaodiao <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(node-host): sync rawCommand with hardened argv after executable path pinning (#33137) Merged via squash. Prepared head SHA: a7987905f7ad6cf5fee286ffa81ceaad8297174f Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * Agents: add generic poll-vote action support * fix(ollama): pass provider headers to Ollama stream function (#24285) createOllamaStreamFn() only accepted baseUrl, ignoring custom headers configured in models.providers.<provider>.headers. This caused 403 errors when Ollama endpoints are behind reverse proxies that require auth headers (e.g. X-OLLAMA-KEY via HAProxy). Add optional defaultHeaders parameter to createOllamaStreamFn() and merge them into every fetch request. Provider headers from config are now passed through at the call site in the embedded runner. Fixes #24285 * test(ollama): add default header precedence coverage * chore(changelog): add PR entry openclaw#24337 thanks @echoVic * Outbound: allow text-only plugin adapters * Outbound: avoid empty multi-media fallback sends * chore(changelog): align outbound adapter entry openclaw#32788 thanks @liuxiaopai-ai * fix(outbound): fail media-only text-only adapter fallback * chore(changelog): clarify outbound media-only fallback openclaw#32788 thanks @liuxiaopai-ai * fix(review): enforce behavioral sweep validation * Fix gateway restart false timeouts on Debian/systemd (#34874) * daemon(systemd): target sudo caller user scope * test(systemd): cover sudo user scope commands * infra(ports): fall back to ss when lsof missing * test(ports): verify ss fallback listener detection * cli(gateway): use probe fallback for restart health * test(gateway): cover restart-health probe fallback * Compaction/Safeguard: require structured summary headings (#25555) Merged via squash. Prepared head SHA: 0b1df34806a7b788261290be55760fd89220de53 Co-authored-by: rodrigouroz <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * Fix Linux daemon install checks when systemd user bus env is missing (#34884) * daemon(systemd): fall back to machine user scope when user bus is missing * test(systemd): cover machine scope fallback for user-bus errors * test(systemd): reset execFile mock state across cases * test(systemd): make machine-user fallback assertion portable * fix(daemon): keep root sudo path on direct user scope * test(systemd): cover sudo root user-scope behavior * ci: use resolvable bun version in setup-node-env * agents: preserve totalTokens on request failure instead of using contextWindow (#34275) Merged via squash. Prepared head SHA: f9d111d0a79a07815d476356e98a28df3a0000ba Co-authored-by: RealKai42 <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix: align AGENTS.md template section names with post-compaction extraction (#25029) (#25098) Merged via squash. Prepared head SHA: 8cd6cc8049aab5a94d8a9d5fb08f2e792c4ac5fd Co-authored-by: echoVic <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * Changelog: add daemon systemd user-bus fallback entry (#34884) * Changelog: add gateway restart health entry (#34874) * fix: finalize spanish locale support * fix: add spanish locale support (#35038) (thanks @DaoPromociones) * fix(deps): patch hono transitive audit vulnerabilities * fix(security): avoid prototype-chain account path checks (#34982) Merged via squash. Prepared head SHA: f89cc6a649959997fe1dec1e1c1bff9a61b2de98 Co-authored-by: HOYALIM <[email protected]> Co-authored-by: dvrshil <[email protected]> Reviewed-by: @dvrshil * fix(deps): bump tar to 7.5.10 * docs(changelog): document dependency security fixes * fix: restore auto-reply system events timeline (#34794) (thanks @anisoptera) (#34794) Co-authored-by: Ayaan Zaidi <[email protected]> * fix(feishu): comprehensive reply mechanism — outbound replyToId forwarding + topic-aware reply targeting (#33789) * fix(feishu): comprehensive reply mechanism fix — outbound replyToId forwarding + topic-aware reply targeting - Forward replyToId from ChannelOutboundContext through sendText/sendMedia to sendMessageFeishu/sendMarkdownCardFeishu/sendMediaFeishu, enabling reply-to-message via the message tool. - Fix group reply targeting: use ctx.messageId (triggering message) in normal groups to prevent silent topic thread creation (#32980). Preserve ctx.rootId targeting for topic-mode groups (group_topic/group_topic_sender) and groups with explicit replyInThread config. - Add regression tests for both fixes. Fixes #32980 Fixes #32958 Related #19784 * fix: normalize Feishu delivery.to before comparing with messaging tool targets - Add normalizeDeliveryTarget helper to strip user:/chat: prefixes for Feishu - Apply normalization in matchesMessagingToolDeliveryTarget before comparison - This ensures cron duplicate suppression works when session uses prefixed targets (user:ou_xxx) but messaging tool extract uses normalized bare IDs (ou_xxx) Fixes review comment on PR #32755 (cherry picked from commit fc20106f16ccc88a5f02e58922bb7b7999fe9dcd) * fix(feishu): catch thrown SDK errors for withdrawn reply targets The Feishu Lark SDK can throw exceptions (SDK errors with .code or AxiosErrors with .response.data.code) for withdrawn/deleted reply targets, in addition to returning error codes in the response object. Wrap reply calls in sendMessageFeishu and sendCardFeishu with try-catch to handle thrown withdrawn/not-found errors (230011, 231003) and fall back to client.im.message.create, matching the existing response-level fallback behavior. Also extract sendFallbackDirect helper to deduplicate the direct-send fallback block across both functions. Closes #33496 (cherry picked from commit ad0901aec103a2c52f186686cfaf5f8ba54b4a48) * feishu: forward outbound reply target context (cherry picked from commit c129a691fcf552a1cebe1e8a22ea8611ffc3b377) * feishu extension: tighten reply target fallback semantics (cherry picked from commit f85ec610f267020b66713c09e648ec004b2e26f1) * fix(feishu): align synthesized fallback typing and changelog attribution * test(feishu): cover group_topic_sender reply targeting --------- Co-authored-by: Xu Zimo <[email protected]> Co-authored-by: Munem Hashmi <[email protected]> Co-authored-by: bmendonca3 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(feishu): use msg_type media for mp4 video (fixes #33674) (#33720) * fix(feishu): use msg_type media for mp4 video (fixes #33674) * Feishu: harden streaming merge semantics and final reply dedupe Use explicit streaming update semantics in the Feishu reply dispatcher: treat onPartialReply payloads as snapshot updates and block fallback payloads as delta chunks, then merge final text with the shared overlap-aware mergeStreamingText helper before closing the stream. Prevent duplicate final text delivery within the same dispatch cycle, and add regression tests covering overlap snapshot merge, duplicate final suppression, and block-as-delta behavior to guard against repeated/truncated output. * fix(feishu): prefer message.reply for streaming cards in topic threads * fix: reduce Feishu streaming card print_step to avoid duplicate rendering Fixes openclaw/openclaw#33751 * Feishu: preserve media sends on duplicate finals and add media synthesis changelog * Feishu: only dedupe exact duplicate final replies * Feishu: use scoped plugin-sdk import in streaming-card tests --------- Co-authored-by: 倪汉杰0668001185 <[email protected]> Co-authored-by: zhengquanliu <[email protected]> Co-authored-by: nick <[email protected]> Co-authored-by: linhey <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(agents): bypass pendingDescendantRuns guard for cron announce delivery (#35185) * fix(agents): bypass pendingDescendantRuns guard for cron announce delivery Standalone cron job completions were blocked from direct channel delivery when the cron run had spawned subagents that were still registered as pending. The pendingDescendantRuns guard exists for live orchestration coordination and should not apply to fire-and-forget cron announce sends. Thread the announceType through the delivery chain and skip both the child-descendant and requester-descendant pending-run guards when the announce originates from a cron job. Closes #34966 * fix: ensure outbound session entry for cron announce with named agents (#32432) Named agents may not have a session entry for their delivery target, causing the announce flow to silently fail (delivered=false, no error). Two fixes: 1. Call ensureOutboundSessionEntry when resolving the cron announce session key so downstream delivery can find channel metadata. 2. Fall back to direct outbound delivery when announce delivery fails to ensure cron output reaches the target channel. Closes #32432 Co-Authored-By: Claude Opus 4.6 <[email protected]> * fix: guard announce direct-delivery fallback against suppression leaks (#32432) The `!delivered` fallback condition was too broad — it caught intentional suppressions (active subagents, interim messages, SILENT_REPLY_TOKEN) in addition to actual announce delivery failures. Add an `announceDeliveryWasAttempted` flag so the direct-delivery fallback only fires when `runSubagentAnnounceFlow` was actually called and failed. Also remove the redundant `if (route)` guard in `resolveCronAnnounceSessionKey` since `resolved` being truthy guarantees `route` is non-null. Co-Authored-By: Claude Opus 4.6 <[email protected]> * fix(cron): harden announce synthesis follow-ups --------- Co-authored-by: scoootscooob <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * Feishu: harden streaming merge semantics and final reply dedupe (#33245) * Feishu: close duplicate final gap and cover routing precedence * Feishu: resolve reviewer duplicate-final and routing feedback * Feishu: tighten streaming send-mode option typing * Feishu: fix reverse-overlap streaming merge ordering * Feishu: align streaming final dedupe test expectation * Feishu: allow distinct streaming finals while deduping repeats --------- Co-authored-by: Tak Hoffman <[email protected]> * fix: cron backup should preserve pre-edit snapshot (#35195) (#35234) * fix(cron): avoid overwriting .bak during normalization Fixes openclaw/openclaw#35195 * test(cron): preserve pre-edit bak snapshot in normalization path --------- Co-authored-by: 0xsline <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(cron): stabilize restart catch-up replay semantics (#35351) * Cron: stabilize restart catch-up replay semantics * Cron: respect backoff in startup missed-run replay * cron: narrow startup replay backoff guard (#35391) * cron: unify stale-run recovery and preserve manual-run every anchors (#35363) * cron: unify stale-run recovery and preserve manual every anchors * cron: address unresolved review threads on recovery paths * cron: remove duplicate timestamp helper after rebase * refactor(telegram): remove unused webhook callback helper (#27816) * fix(pr): make review claim step required * fix(skills): deduplicate slash commands by skillName across all interfaces Move skill-command deduplication by skillName from the Discord-only `dedupeSkillCommandsForDiscord` into `listSkillCommandsForAgents` so every interface (TUI, Slack, text) consistently sees a clean command list without platform-specific workarounds. When multiple agents share a skill with the same name the old code emitted `github` + `github_2` and relied on Discord to collapse them. Now `listSkillCommandsForAgents` returns only the first registration per skillName, and the Discord-specific wrapper is removed. Co-Authored-By: Claude Sonnet 4.6 <[email protected]> * style: fix formatting in skill-commands.test.ts and provider.ts Co-Authored-By: Claude Sonnet 4.6 <[email protected]> * style(skills): align formatting cleanup for dedupe changes * chore(changelog): add dedupe note openclaw#27521 thanks @shivama205 * fix(agents): detect Venice provider proxying xAI/Grok models for schema cleaning (#35355) Merged via squash. Prepared head SHA: 8bfdec257bb6a6025cb69a0a213a433da32b15db Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(agents): decode HTML entities in xAI/Grok tool call arguments (#35276) Merged via squash. Prepared head SHA: c4445d2938898ded9c046614f9315dbda65ec573 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(agents): guard promoteThinkingTagsToBlocks against malformed content entries (#35143) Merged via squash. Prepared head SHA: 3971122f5fd27c66c8c9c5ce783f00e113b1f47b Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(web-ui): render Accounts schema node properly (#35380) Co-authored-by: stakeswky <[email protected]> Co-authored-by: liuxiaopai-ai <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(agents): guard context pruning against malformed thinking blocks (#35146) Merged via squash. Prepared head SHA: a196a565b1b8e806ffbf85172bcf1128796b45a2 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(gateway): prevent internal route leakage in chat.send Synthesis of routing fixes from #35321, #34635, and #35356 for internal-client reply safety. - Require explicit `deliver: true` before inheriting any external delivery route. - Keep webchat/TUI/UI-origin traffic on internal routing by default. - Allow configured-main session inheritance only for non-Webchat/UI clients, and honor `session.mainKey`. - Add regression tests for UI no-inherit, configured-main CLI inherit, and deliver-flag behavior. Co-authored-by: alexyyyander <[email protected]> Co-authored-by: Octane0411 <[email protected]> Co-authored-by: Linux2010 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(gateway): pass actual version to Control UI client instead of dev (#35230) * fix(gateway): pass actual version to Control UI client instead of "dev" The GatewayClient, CLI WS client, and browser Control UI all sent "dev" as their clientVersion during handshake, making it impossible to distinguish builds in gateway logs and health snapshots. - GatewayClient and CLI WS client now use the resolved VERSION constant - Control UI reads serverVersion from the bootstrap endpoint and forwards it when connecting - Bootstrap contract extended with serverVersion field Closes #35209 * Gateway: fix control-ui version version-reporting consistency * Control UI: guard deferred bootstrap connect after disconnect * fix(ui): accept same-origin http and relative gateway URLs for client version --------- Co-authored-by: Tak Hoffman <[email protected]> * chore(pr): enforce changelog placement and reduce merge sync churn * TTS: add baseUrl support to OpenAI TTS config (#34321) Merged via squash. Prepared head SHA: e9a10cf81d2021cf81091dfa81e13ffdbb6a540a Co-authored-by: RealKai42 <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * ACP: add persistent Discord channel and Telegram topic bindings (#34873) * docs: add ACP persistent binding experiment plan * docs: align ACP persistent binding spec to channel-local config * docs: scope Telegram ACP bindings to forum topics only * docs: lock bound /new and /reset behavior to in-place ACP reset * ACP: add persistent discord/telegram conversation bindings * ACP: fix persistent binding reuse and discord thread parent context * docs: document channel-specific persistent ACP bindings * ACP: split persistent bindings and share conversation id helpers * ACP: defer configured binding init until preflight passes * ACP: fix discord thread parent fallback and explicit disable inheritance * ACP: keep bound /new and /reset in-place * ACP: honor configured bindings in native command flows * ACP: avoid configured fallback after runtime bind failure * docs: refine ACP bindings experiment config examples * acp: cut over to typed top-level persistent bindings * ACP bindings: harden reset recovery and native command auth * Docs: add ACP bound command auth proposal * Tests: normalize i18n registry zh-CN assertion encoding * ACP bindings: address review findings for reset and fallback routing * ACP reset: gate hooks on success and preserve /new arguments * ACP bindings: fix auth and binding-priority review findings * Telegram ACP: gate ensure on auth and accepted messages * ACP bindings: fix session-key precedence and unavailable handling * ACP reset/native commands: honor fallback targets and abort on bootstrap failure * Config schema: validate ACP binding channel and Telegram topic IDs * Discord ACP: apply configured DM bindings to native commands * ACP reset tails: dispatch through ACP after command handling * ACP tails/native reset auth: fix target dispatch and restore full auth * ACP reset detection: fallback to active ACP keys for DM contexts * Tests: type runTurn mock input in ACP dispatch test * ACP: dedup binding route bootstrap and reset target resolution * reply: align ACP reset hooks with bound session key * docs: replace personal discord ids with placeholders * fix: add changelog entry for ACP persistent bindings (#34873) (thanks @dutifulbob) --------- Co-authored-by: Onur <[email protected]> * docs(telegram): recommend allowlist for single-user DM policy (#34841) * docs(telegram): recommend allowlist for single-user bots * docs(telegram): condense single-user allowlist note --------- Co-authored-by: echoVic <[email protected]> * fix(feishu): check response.ok before calling response.json() in streaming card (#35628) Merged via squash. Prepared head SHA: 62c3fec80d97cea9be344c0bef5358a0a5dc5560 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * Mattermost: honor onmessage mention override and add gating diagnostics tests (#27160) Merged via squash. Prepared head SHA: 6cefb1d5bf3d6dfcec36c1cee3f9ea887f10c890 Co-authored-by: turian <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Reviewed-by: @mukhtharcm * fix(subagents): strip leaked [[reply_to]] tags from completion announces (#34503) * fix(subagents): strip reply tags from completion delivery text * test(subagents): cover reply-tag stripping in cron completion sends * changelog: note iMessage reply-tag stripping in completion announces * Update CHANGELOG.md * Apply suggestion from @greptile-apps[bot] Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> --------- Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix(cron): restore direct fallback after announce failure in best-effort mode (openclaw#36177) Verified: - pnpm build - pnpm check (fails on pre-existing origin/main lint debt in extensions/mattermost imports) - pnpm test:macmini Co-authored-by: Tak Hoffman <[email protected]> * test(cron): add cross-channel announce fallback regression coverage (openclaw#36197) Verified: - pnpm install --frozen-lockfile - pnpm build - pnpm check (fails on pre-existing origin/main lint debt in extensions/mattermost imports) - pnpm test:macmini Co-authored-by: Tak Hoffman <[email protected]> * feat(mattermost): add interactive buttons support (#19957) Merged via squash. Prepared head SHA: 8a25e608729d0b9fd07bb0ee4219d199d9796dbe Co-authored-by: tonydehnke <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Reviewed-by: @mukhtharcm * fix(browser): remove deprecated --disable-blink-features=AutomationControlled flag - Removes OpenClaw's default `--disable-blink-features=AutomationControlled` Chrome launch switch to avoid unsupported-flag warnings in newer Chrome (#35721). - Preserves compatibility for older Chrome via `browser.extraArgs` override behavior (source analysis: #35770, #35728, #35727, #35885). - Synthesis attribution: thanks @Sid-Qin, @kevinWangSheng, @ningding97, @Naylenv, @clawbie. Source PR refs: #35734, #35770, #35728, #35727, #35885 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: kevinWangSheng <[email protected]> Co-authored-by: ningding97 <[email protected]> Co-authored-by: Naylenv <[email protected]> Co-authored-by: clawbie <[email protected]> Co-authored-by: Takhoffman <[email protected]> * fix(feishu): add HTTP timeout to prevent per-chat queue deadlocks (#36430) When the Feishu API hangs or responds slowly, the sendChain never settles, causing the per-chat queue to remain in a processing state forever and blocking all subsequent messages in that thread. This adds a 30-second default timeout to all Feishu HTTP requests by providing a timeout-aware httpInstance to the Lark SDK client. Closes #36412 Co-authored-by: Ayane <[email protected]> * fix(feishu): use probed botName for mention checks (#36391) * Feishu: honor bot mentions by ID despite aliases (Fixes #36317) (#36333) * Mattermost: switch plugin-sdk imports to scoped subpaths (openclaw#36480) Verified: - pnpm build - pnpm check - pnpm test:macmini Co-authored-by: Takhoffman <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(feishu): accept groupPolicy "allowall" as alias for "open" (#36358) * fix(feishu): accept groupPolicy "allowall" as alias for "open" When users configure groupPolicy: "allowall" in Feishu channel config, the Zod schema rejects the value and the runtime policy check falls through to the allowlist path. With an empty allowFrom array, all group messages are silently dropped despite the intended "allow all" semantics. Accept "allowall" at the schema level (transform to "open") and add a runtime guard in isFeishuGroupAllowed so the value is handled even if it bypasses schema validation. Closes #36312 Made-with: Cursor * Feishu: tighten allowall alias handling and coverage --------- Co-authored-by: Tak Hoffman <[email protected]> * synthesis: fix Feishu group mention slash parsing ## Summary\n\nFeishu group slash command parsing is fixed for mentions and command probes across authorization paths.\n\nThis includes:\n- Normalizing bot mention text in group context for reliable slash detection in message parsing.\n- Adding command-probe normalization for group slash invocations.\n\nCo-authored-by: Sid Qin <[email protected]>\nCo-authored-by: Tak Hoffman <[email protected]> * Feishu: normalize group slash command probing - Feishu/group slash command detection: normalize group mention wrappers before command-authorization probing so mention-prefixed commands are recognized in group routing.\n- Source PR: #36011\n- Contributor: @liuxiaopai-ai\n\nCo-authored-by: Tak Hoffman <[email protected]>\nCo-authored-by: liuxiaopai-ai <[email protected]> * add prependSystemContext and appendSystemContext to before_prompt_build (fixes #35131) (#35177) Merged via squash. Prepared head SHA: d9a2869ad69db9449336a2e2846bd9de0e647ac6 Co-authored-by: maweibin <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(feishu): avoid media regressions from global HTTP timeout (#36500) * fix(feishu): avoid media regressions from global http timeout * fix(feishu): source HTTP timeout from config * fix(feishu): apply media timeout override to image uploads * fix(feishu): invalidate cached client when timeout changes * fix(feishu): clamp timeout values and cover image download * Gateway: add SecretRef support for gateway.auth.token with auth-mode guardrails (#35094) * fix(embedded): classify model_context_window_exceeded as context overflow, trigger compaction (#35934) Merged via squash. Prepared head SHA: 20fa77289c80b2807a6779a3df70440242bc18ca Co-authored-by: RealKai42 <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(agents): skip compaction API call when session has no real messages (#36451) Merged via squash. Prepared head SHA: 52dd6317895c7bd10855d2bd7dbbfc2f5279b68e Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(ui): catch marked.js parse errors to prevent Control UI crash (#36445) - Prevent Control UI session render crashes when `marked.parse()` encounters pathological recursive markdown by safely falling back to escaped `<pre>` output. - Tighten markdown fallback regression coverage and keep changelog attribution in sync for this crash-hardening path. Co-authored-by: Bin Deng <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(session): archive old transcript on daily/scheduled reset to prevent orphaned files (#35493) Merged via squash. Prepared head SHA: 0d95549d752adecfc0b08d5cd55a8b8c75e264fe Co-authored-by: byungsker <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(agents): set preserveSignatures to isAnthropic in resolveTranscriptPolicy (#32813) Merged via squash. Prepared head SHA: f522d21ca59a42abac554435a0aa646f6a34698d Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix: avoid false global rate-limit classification from generic cooldown text (#32972) Merged via squash. Prepared head SHA: 813c16f5afce415da130a917d9ce9f968912b477 Co-authored-by: stakeswky <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * refactor(agents): share failover HTTP status classification (#36615) * fix(agents): classify transient failover statuses consistently * fix(agents): preserve legacy failover status mapping * fix(failover): narrow service-unavailable to require overload indicator (#32828) (#36646) Merged via squash. Prepared head SHA: 46fb4306127972d7635f371fd9029fbb9baff236 Co-authored-by: jnMetaCode <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * Compaction/Safeguard: add summary quality audit retries (#25556) Merged via squash. Prepared head SHA: be473efd1635616ebbae6e649d542ed50b4a827f Co-authored-by: rodrigouroz <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * test(agents): add provider-backed failover regressions (#36735) * test(agents): add provider-backed failover fixtures * test(agents): cover more provider error docs * test(agents): tighten provider doc fixtures * Docs: add Slack typing reaction fallback * Docs: update gateway config reference for Slack and TTS * Docs: clarify OpenAI-compatible TTS endpoints * Docs: document Control UI locale support * Docs: cover heartbeat, cron, and plugin route updates * fix(ui): bump dompurify to 3.3.2 (#36781) * UI: bump dompurify to 3.3.2 * Deps: refresh dompurify lockfile * UI: hoist lifecycle connect test mocks (#36788) * fix(agents): classify insufficient_quota 400s as billing (#36783) * feat: append UTC time alongside local time in shared Current time lines (#32423) Merged via squash. Prepared head SHA: 9e8ec13933b5317e7cff3f0bc048de515826c31a Co-authored-by: jriff <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * fix(auth): grant senderIsOwner for internal channels with operator.admin scope (openclaw#35704) Verified: - pnpm install --frozen-lockfile - pnpm build - pnpm check - pnpm test:macmini Co-authored-by: Naylenv <[email protected]> Co-authored-by: Octane0411 <[email protected]> Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(config): prevent RangeError in merged schema cache key generation Fix merged schema cache key generation for high-cardinality plugin/channel metadata by hashing incrementally instead of serializing one large aggregate string. Includes changelog entry for the user-visible regression fix. Co-authored-by: Tak Hoffman <[email protected]> Co-authored-by: Bill <[email protected]> * fix(slack): propagate mediaLocalRoots through Slack send path Restore Slack local file upload parity with CVE-era local media allowlist enforcement by threading `mediaLocalRoots` through the Slack send call chain. - pass `ctx.mediaLocalRoots` from Slack channel action adapter into `handleSlackAction` - add and forward `mediaLocalRoots` in Slack action context/send path - pass `mediaLocalRoots` into `sendMessageSlack` for upload allowlist enforcement - add changelog entry with attribution for this behavior fix Co-authored-by: 2233admin <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(slack): preserve dedupe while recovering dropped app_mention (#34937) This PR fixes Slack mention loss without reintroducing duplicate dispatches. - Preserve seen-message dedupe at ingress to prevent duplicate processing. - Allow a one-time app_mention retry only when the paired message event was previously dropped before dispatch. - Add targeted race tests for both recovery and duplicate-prevention paths. Co-authored-by: littleben <[email protected]> Co-authored-by: OpenClaw Agent <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * README: add algal to contributors list (#2046) * fix: decouple Discord inbound worker timeout from listener timeout (#36602) (thanks @dutifulbob) (#36602) Co-authored-by: Onur Solmaz <[email protected]> * plugins: enforce prompt hook policy with runtime validation (#36567) Merged via squash. Prepared head SHA: 6b9d883b6ae33628235fb02ce39c0d0f46a065bb Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(memory): avoid destructive qmd collection rebinds * Harden Telegram poll gating and schema consistency (#36547) Merged via squash. Prepared head SHA: f77824419e3d166f727474a9953a063a2b4547f2 Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(browser): close tracked tabs on session cleanup (#36666) * Diffs: restore system prompt guidance (#36904) Merged via squash. Prepared head SHA: 1b3be3c87957c068473d5c86b9efba4a1a8503f2 Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(routing): avoid full binding rescans in resolveAgentRoute (#36915) * fix(gateway): honor insecure ws override for remote hostnames * fix(llm-task): load runEmbeddedPiAgent from dist/extensionAPI in installs * fix(auth): harden openai-codex oauth login path * feat(telegram/acp): Topic Binding, Pin Binding Message, Fix Spawn Param Parsing (#36683) * fix(acp): normalize unicode flags and Telegram topic binding * feat(telegram/acp): restore topic-bound ACP and session bindings * fix(acpx): clarify permission-denied guidance * feat(telegram/acp): pin spawn bind notice in topics * docs(telegram): document ACP topic thread binding behavior * refactor(reply): share Telegram conversation-id resolver * fix(telegram/acp): preserve bound session routing semantics * fix(telegram): respect binding persistence and expiry reporting * refactor(telegram): simplify binding lifecycle persistence * fix(telegram): bind acp spawns in direct messages * fix: document telegram ACP topic binding changelog (#36683) (thanks @huntharo) --------- Co-authored-by: Onur <[email protected]> * fix(gateway): preserve streamed prefixes across tool boundaries * fix(tui): prevent stale model indicator after /model * Memory: handle SecretRef keys in doctor embeddings (#36835) Merged via squash. Prepared head SHA: c1a3d0caae60115d886e8bfc9983c9533c773f04 Co-authored-by: joshavant <[email protected]> Co-authored-by: joshavant <[email protected]> Reviewed-by: @joshavant * fix(openai-codex): request required oauth api scopes (#24720) * fix(memory-flush): ban timestamped variant files in default flush prompt (#34951) Merged via squash. Prepared head SHA: efadda4988b460e6da07be72994d4951d64239d0 Co-authored-by: zerone0x <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(tui): render final event error when assistant output is empty (#14687) * feat(agents): flush reply pipeline before compaction wait (#35489) Merged via squash. Prepared head SHA: 7dbbcc510b74b0e8d35eb750d24575e34b5d769a Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(secrets): harden api key normalization for ByteString headers * fix(slack): remove double mrkdwn conversion in native streaming path Remove redundant text normalization from Slack native streaming markdown_text flow so Markdown formatting is preserved. Synthesis context: overlaps reviewed from #34931, #34759, #34716, #34682, #34814. Co-authored-by: littleben <[email protected]> Co-authored-by: dunamismax <[email protected]> Co-authored-by: Octane <[email protected]> Co-authored-by: Mitsuyuki Osabe <[email protected]> Co-authored-by: Kai <[email protected]> Co-authored-by: OpenClaw Agent <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(kimi-coding): normalize anthropic tool payload format * fix(slack): thread channel ID through inbound context for reactions (#34831) Slack reaction/thread context routing fixes via canonical synthesis of #34831. Co-authored-by: Tak <[email protected]> * fix(heartbeat): pin HEARTBEAT.md reads to workspace path * fix(subagents): recover announce cleanup after kill/complete race * feat(hooks): emit compaction lifecycle hooks (#16788) * fix(auth): harden openai-codex oauth refresh fallback * fix(subagents): announce delivery with descendant gating, frozen result refresh, and cron retry (#35080) Thanks @tyler6204 * fix(agents): avoid synthetic tool-result writes on idle-timeout cleanup * fix(agent): harden undici stream timeouts for long openai-completions runs * fix(slack): record app_mention retry key before dedupe check (#37033) - Prime app_mention retry allowance before dedupe so near-simultaneous message/app_mention races do not drop valid mentions. - Prevent duplicate dispatch when app_mention wins the race and message prepare later succeeds. - Prune dispatched mention keys and add regression coverage for both dropped and successful in-flight message outcomes. Co-authored-by: Tak Hoffman <[email protected]> * fix(agents): honor explicit rate-limit cooldown probes in fallback runs * fix(agents): allow configured ollama endpoints without dummy api keys * fix(memory): recover qmd updates from duplicate document constraints * Doctor: warn on implicit heartbeat directPolicy (#36789) * Changelog: note heartbeat directPolicy doctor warning * Tests: cover heartbeat directPolicy doctor warning * Doctor: warn on implicit heartbeat directPolicy * Tests: cover per-agent heartbeat directPolicy warning * Update CHANGELOG.md * Plugins: clarify registerHttpHandler migration errors (#36794) * Changelog: note plugin HTTP route migration diagnostics * Tests: cover registerHttpHandler migration diagnostics * Plugins: clarify registerHttpHandler migration errors * Tests: cover registerHttpHandler diagnostic edge cases * Plugins: tighten registerHttpHandler migration hint * fix(memory): repair qmd collection name conflicts during ensure * fix(memory): handle qmd search results without docid * Plugins: avoid false integrity drift prompts on unpinned updates (#37179) * Plugins: skip drift prompts for unpinned updates * Plugins: cover unpinned integrity update behavior * Changelog: add #37179 release note * Delete changelog/fragments directory * Update CHANGELOG.md * fix(whatsapp): remove implicit [openclaw] self-chat prefix * fix: remove config.schema from agent gateway tool (#7382) Merged via squash. Prepared head SHA: f34a7780690a941936b31899e2d096b8a07f4afc Co-authored-by: kakuteki <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * feat(openai): add gpt-5.4 support for API and Codex OAuth (#36590) * feat(openai): add gpt-5.4 support and priority processing * feat(openai-codex): add gpt-5.4 oauth support * fix(openai): preserve provider overrides in gpt-5.4 fallback * fix(openai-codex): keep xhigh for gpt-5.4 default * fix(models): preserve configured overrides in list output * fix(models): close gpt-5.4 integration gaps * fix(openai): scope service tier to public api * fix(openai): complete prep followups for gpt-5.4 support (#36590) (thanks @dorukardahan) --------- Co-authored-by: Tyler Yust <[email protected]> * fix(tui): preserve credential-like tokens in render sanitization * CLI: make read-only SecretRef status flows degrade safely (#37023) * CLI: add read-only SecretRef inspection * CLI: fix read-only SecretRef status regressions * CLI: preserve read-only SecretRef status fallbacks * Docs: document read-only channel inspection hook * CLI: preserve audit coverage for read-only SecretRefs * CLI: fix read-only status account selection * CLI: fix targeted gateway fallback analysis * CLI: fix Slack HTTP read-only inspection * CLI: align audit credential status checks * CLI: restore Telegram read-only fallback semantics * chore(changelog): update for #37023 Signed-off-by: joshavant <[email protected]> * fix(agents): disable usage streaming chunks on non-native openai-completions * feat(nano-banana-pro): add --aspect-ratio flag to generate_image.py (#28159) * feat(nano-banana-pro): add --aspect-ratio flag to generate_image.py * Nano Banana: allow all supported aspect ratios * Docs: expand nano banana aspect ratio options --------- Co-authored-by: Vincent Koc <[email protected]> * fix(gateway): support image_url in OpenAI chat completions (#34068) * fix(gateway): parse image_url in openai chat completions * test(gateway): cover openai chat completions image_url flows * docs(changelog): note openai image_url chat completions fix (#17685) * fix(gateway): harden openai image_url parsing and limits * test(gateway): add openai image_url regression coverage * docs(changelog): expand #17685 openai chat completions note * Gateway: make OpenAI image_url URL fetch opt-in and configurable * Diagnostics: redact image base64 payload data in trace logs * Changelog: note OpenAI image_url hardening follow-ups * Gateway: enforce OpenAI image_url total budget incrementally * Gateway: scope OpenAI image_url extraction to the active turn * Update CHANGELOG.md * fix(agents): avoid xAI web_search tool-name collisions * fix: clear Telegram DM draft after materialize (#36746) (thanks @joelnishanth) * Fix Control UI duplicate iMessage replies for internal webchat turns (#36151) * Auto-reply: avoid routing external replies from internal webchat turns * Auto-reply tests: cover internal webchat non-routing with external origin metadata * Changelog: add Control UI iMessage duplicate-reply fix note * Auto-reply context: track explicit deliver routes * Gateway chat: mark explicit external deliver routes in context * Auto-reply: preserve explicit deliver routes for internal webchat turns * Auto-reply tests: cover explicit deliver routes from internal webchat turns * Gateway chat tests: assert explicit deliver route context tagging * fix: enforce 600 perms for cron store and run logs (#36078) * fix: enforce secure permissions for cron store and run logs * fix(cron): enforce dir perms and gate posix tests on windows * Cron store tests: cover existing directory permission hardening * Cron run-log tests: cover existing directory permission hardening * Changelog: note cron file permission hardening --------- Co-authored-by: linhey <[email protected]> Co-authored-by: Vincent Koc <[email protected]> * fix(tui): accept canonical session-key aliases in chat event routing * Gateway: normalize OpenAI stream chunk text * Gateway: coerce chat deliverable route boolean * fix(web_search): align brave language codes with API * Respect source channel for agent event surfacing (#36030) * fix(session): prefer webchat routes for direct ui turns (#37135) * Gateway: discriminate input sources * Cron: migrate legacy provider delivery hints * Cron: stabilize runs-one-shot migration tests * fix(memory): retry mcporter after Windows EINVAL spawn * fix(onboarding): guard daemon status probe on headless linux * Gateway: add path-scoped config schema lookup (#37266) Merged via squash. Prepared head SHA: 0c4d187f6fb66f2799d4047585d6368e433c883a Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * docs(changelog): add pr entry * fix(ci): restore protocol and schema checks (#37470) * Fix failover for zhipuai 1310 Weekly/Monthly Limit Exhausted (#33813) Merged via squash. Prepared head SHA: 3dc441e58de48913720cf7b6137fa761758d8344 Co-authored-by: zhouhe-xydt <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * fix(openai-codex-oauth): stop mutating authorize url scopes * Update CHANGELOG.md * fix(auth): remove bogus codex oauth responses probe * docs(changelog): fold codex oauth fix notes * docs(changelog): add codex oauth pr reference (#37558) * fix(security:PLA-697): block local secret scratch files from commits * feat: add dev EKS deployment for openclaw SRE container * fix: stabilize slack socket mode in dev eks runtime * fix(eks): reuse monitoring incident auth for openclaw-sre * feat(deploy): add grafana env guards and prod deploy wrapper * fix(security): redact tool outputs and enforce secret-safe runtime defaults * feat(sre): harden heartbeat routing and enrich triage signals * docs: add SRE hybrid intelligence design Three-layer architecture to improve bot reasoning quality: - Layer 1: Service knowledge (auto-discovery + overlays + incident memory) - Layer 2: Multi-stage reasoning chain (triage → hypothesize → causal chain → action plan → cross-review) - Layer 3: Incident learning loop (structured cards, overlay suggestions, feedback signals) Co-Authored-By: Claude Opus 4.6 <[email protected]> * feat(sandbox): add boundary cli to common runtime image * docs(sre): finalize hybrid intelligence design v19 after 18 Codex review rounds Iteratively hardened the design through 18 adversarial Codex review rounds, resolving 60+ findings (15+ CRITICAL, 45+ HIGH). Key fixes include: - Decoupled incident_id (pre-Step11 immutable) from card_id (LLM-derived) - Unified evidence/memory sanitization via shared _strip_instruction_tokens - Fixed rerun interval default (3600s > heartbeat 1800s) to prevent every-cycle re-runs - Added Phase 2 cross-review gap note and dual-column depth table - Overlay suggestion idempotency via deterministic suggestion_key + upsert - Decoupled chain timeout from legacy budget check - Added predicate alignment docs for L3 pre-check vs L3 gate - Adopted dynamic evidence_completeness denominator matching existing code - Added scope note clarifying design vs current implementation Co-Authored-By: Claude Opus 4.6 <[email protected]> * feat(sre): complete hybrid intelligence rollout and bot hardening * fix(sre): harden slack reasoning, dual chain, and deploy auth/runtime * ci: add ecr release pipeline and infra-helm image update trigger * ci: switch workflows to github-hosted runners * ci: allow manual ecr env target and use updater v4 * ci: remove windows jobs from workflow * chore(pla-678): update fallback auth fixture and plan docs * fix(pla-678): preserve control-ui host-header fallback in dev config * refactor(pla-678): remove local sre chart and require infra-helm * refactor(pla-678): remove direct eks deploy scripts * ci(pla-678): remove labeler workflow * ci(pla-678): refresh checks after disabling labeler * build: install boundary CLI in SRE runtime image * fix(slack:PLA-678): keep one thread lane across users * fix(sre-skill): enforce live db-query runbook in dev seed * fix(slack): ignore bot thread history when seeding new sessions * fix(slack): force retry on repeated thread questions * fix(sre-skill): require retry on repeated db requests * docs(sre-skill): make retry policy generic across repeated asks * fix(sre:PLA-678): enforce live linear ticket updates * fix(sre:PLA-678): use [PLATFORM] Backlog linear project * fix(sre:PLA-678): label linked linear tickets as openclaw-sre * fix(sre:PLA-678): auto-label linked linear tickets for auto-pr * test(sre:PLA-678): make tracking label test executable * fix(slack:PLA-678): remove status-final completion banner (#16) * fix(slack:PLA-678): enable direct file attachments from agent replies (#17) * fix(slack:PLA-678): remove status-final completion banner * fix(slack): enable direct file attachments from agent replies * refactor(sre:PLA-678): remove local deploy config mirror (#18) * fix: compact oversized auto-generated pr bodies * fix(sre): add built-in linear incident memory provider * feat(sre): add erpc helper with flo secret query * fix(sre): enforce canonical rpc.morpho.dev erpc endpoint * refactor(sre): remove local deploy config mirror * fix(slack:PLA-678): force DM replies into per-message threads (#19) * fix: compact oversized auto-generated pr bodies * fix(sre): add built-in linear incident memory provider * feat(sre): add erpc helper with flo secret query * fix(sre): enforce canonical rpc.morpho.dev erpc endpoint * refactor(sre): remove local deploy config mirror * fix(slack): force threaded replies for direct messages * feat(cron:PLA-678): use conversation history in self-improve runs (#20) * fix(slack:PLA-678): improve progress update readability (#21) * fix(ci:PLA-678): restore main release pipeline health (#22) * fix(ci:PLA-678): restore main release pipeline health * fix(ci:PLA-678): harden auto-response token fallback * fix(ci:PLA-678): pin setup-bun to published release * test(ci:PLA-678): align slack thread session expectations * fix(sre:PLA-678): harden Slack reply delivery and image release flow (#23) * fix(slack:PLA-678): harden thread-scoped reply delivery * fix(slack:PLA-678): satisfy lint after rebase * ci(release:PLA-678): move image release to cached ecr flow * ci(release:PLA-678): use official docker build actions * fix(slack:PLA-722): preserve progress updates during final-answer gating (#24) * fix(sre:PLA-724): restore ECR runtime image toolchain (#25) * fix(failover): classify HTTP 402 as rate_limit when payload indicates usage limit (#30484) (#36802) * fix(failover): classify HTTP 402 as rate_limit when payload indicates usage limit (#30484) Some providers (notably Anthropic Claude Max plan) surface temporary usage/rate-limit failures as HTTP 402 instead of 429. Before this change, all 402s were unconditionally mapped to 'billing', which produced a misleading 'run out of credits' warning for Max plan users who simply hit their usage window. This follows the same pattern introduced for HTTP 400 in #36783: check the error message for an explicit rate-limit signal before falling back to the default status-code classification. - classifyFailoverReasonFromHttpStatus now returns 'rate_limit' for 402 when isRateLimitErrorMessage matches the payload text - Added regression tests covering both the rate-limit and billing paths on 402 * fix: narrow 402 rate-limit matcher to prevent billing misclassification The original implementation used isRateLimitErrorMessage(), which matches phrases like 'quota exceeded' that legitimately appear in billing errors. This commit replaces it with a narrow, 402-specific matcher that requires BOTH retry language (try again/retry/temporary/cooldown) AND limit terminology (usage limit/rate limit/organization usage). Prevents misclassification of errors like: 'HTTP 402: exceeded quota, please add credits' -> billing (not rate_limit) Added regression test for the ambiguous case. --------- Co-authored-by: Val Alexander <[email protected]> * fix(mattermost): allow reachable interaction callback URLs (#37543) Merged via squash. Prepared head SHA: 4d593731be5a5dcbf3106d596b38acfeb8cf0aa8 Co-authored-by: mukhtharcm <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Reviewed-by: @mukhtharcm * fix(cron:PLA-740): audit previous-day self-improve sessions --------- Signed-off-by: joshavant <[email protected]> Co-authored-by: Gustavo Madeira Santana <[email protected]> Co-authored-by: Mariano <[email protected]> Co-authored-by: leepokai <[email protected]> Co-authored-by: Bob <[email protected]> Co-authored-by: Onur <[email protected]> Co-authored-by: Ayaan Zaidi <[email protected]> Co-authored-by: Vincent Koc <[email protected]> Co-authored-by: huangcj <[email protected]> Co-authored-by: gumadeiras <[email protected]> Co-authored-by: Sid <[email protected]> Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Co-authored-by: a <[email protected]> Co-authored-by: echoVic <[email protected]> Co-authored-by: Shakker <[email protected]> Co-authored-by: liuxiaopai-ai <[email protected]> Co-authored-by: Rodrigo Uroz <[email protected]> Co-authored-by: rodrigouroz <[email protected]> Co-authored-by: jalehman <[email protected]> Co-authored-by: Kai <[email protected]> Co-authored-by: RealKai42 <[email protected]> Co-authored-by: 青雲 <[email protected]> Co-authored-by: echoVic <[email protected]> Co-authored-by: Darshil <[email protected]> Co-authored-by: Ho Lim <[email protected]> Co-authored-by: dvrshil <[email protected]> Co-authored-by: Isis Anisoptera <[email protected]> Co-authored-by: Madoka <[email protected]> Co-authored-by: Xu Zimo <[email protected]> Co-authored-by: Munem Hashmi <[email protected]> Co-authored-by: bmendonca3 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> Co-authored-by: Nhj <[email protected]> Co-authored-by: 倪汉杰0668001185 <[email protected]> Co-authored-by: zhengquanliu <[email protected]> Co-authored-by: nick <[email protected]> Co-authored-by: linhey <[email protected]> Co-authored-by: scoootscooob <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> Co-authored-by: rexl2018 <[email protected]> Co-authored-by: sline <[email protected]> Co-authored-by: 0xsline <[email protected]> Co-authored-by: Harold Hunt <[email protected]> Co-authored-by: Shivam <[email protected]> Co-authored-by: 不做了睡大觉 <[email protected]> Co-authored-by: alexyyyander <[email protected]> Co-authored-by: alexyyyander <[email protected]> Co-authored-by: Octane0411 <[email protected]> Co-authored-by: Linux2010 <[email protected]> Co-authored-by: echoVic <[email protected]> Co-authored-by: Joseph Turian <[email protected]> Co-authored-by: turian <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> Co-authored-by: Tony Dehnke <[email protected]> Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: kevinWangSheng <[email protected]> Co-authored-by: ningding97 <[email protected]> Co-authored-by: Naylenv <[email protected]> Co-authored-by: clawbie <[email protected]> Co-authored-by: Takhoffman <[email protected]> Co-authored-by: Ayane <[email protected]> Co-authored-by: Ayane <[email protected]> Co-authored-by: StingNing <[email protected]> Co-authored-by: maweibin <[email protected]> Co-authored-by: maweibin <[email protected]> Co-authored-by: Josh Avant <[email protected]> Co-authored-by: Bin Deng <[email protected]> Co-authored-by: Byungsker <[email protected]> Co-authored-by: altaywtf <[email protected]> Co-authored-by: Altay <[email protected]> Co-authored-by: jiangnan <[email protected]> Co-authored-by: jnMetaCode <[email protected]> Co-authored-by: Jacob Riff <[email protected]> Co-authored-by: jriff <[email protected]> Co-authored-by: Naylenv <[email protected]> Co-authored-by: Bill <[email protected]> Co-authored-by: 2233admin <[email protected]> Co-authored-by: 2233admin <[email protected]> Co-authored-by: littleben <[email protected]> Co-authored-by: littleben <[email protected]> Co-authored-by: OpenClaw Agent <[email protected]> Co-authored-by: Vignesh Natarajan <[email protected]> Co-authored-by: zerone0x <[email protected]> Co-authored-by: zerone0x <[email protected]> Co-authored-by: dunamismax <[email protected]> Co-authored-by: Octane <[email protected]> Co-authored-by: Mitsuyuki Osabe <[email protected]> Co-authored-by: Tak <[email protected]> Co-authored-by: Tyler Yust <[email protected]> Co-authored-by: Hinata Kaga (samon) <[email protected]> Co-authored-by: dorukardahan <[email protected]> Co-authored-by: Tyler Yust <[email protected]> Co-authored-by: Brenner Spear <[email protected]> Co-authored-by: aerelune <[email protected]> Co-authored-by: Frank Yang <[email protected]> Co-authored-by: zhouhe-xydt <[email protected]> Co-authored-by: zhouhe-xydt <[email protected]> Co-authored-by: Vignesh <[email protected]> Co-authored-by: OpenClaw SRE Bot <[email protected]> Co-authored-by: Xinhua Gu <[email protected]> Co-authored-by: Val Alexander <[email protected]>
0x666c6f
added a commit
to 0x666c6f/openclaw
that referenced
this pull request
Mar 26, 2026
…openclaw#31) * Plugins/memory-lancedb: migrate to scoped plugin-sdk imports * Plugins/minimax-portal-auth: migrate to scoped plugin-sdk imports * Plugins/msteams: migrate to scoped plugin-sdk imports * Plugins/nextcloud-talk: migrate to scoped plugin-sdk imports * Plugins/nostr: migrate to scoped plugin-sdk imports * Plugins/open-prose: migrate to scoped plugin-sdk imports * Plugins/phone-control: migrate to scoped plugin-sdk imports * Plugins/qwen-portal-auth: migrate to scoped plugin-sdk imports * Plugins/synology-chat: migrate to scoped plugin-sdk imports * Plugins/talk-voice: migrate to scoped plugin-sdk imports * Plugins/test-utils: migrate to scoped plugin-sdk imports * Plugins/thread-ownership: migrate to scoped plugin-sdk imports * Plugins/tlon: migrate to scoped plugin-sdk imports * Plugins/twitch: migrate to scoped plugin-sdk imports * Plugins/voice-call: migrate to scoped plugin-sdk imports * Plugins/whatsapp: migrate to scoped plugin-sdk imports * Plugins/zalo: migrate to scoped plugin-sdk imports * Plugins/zalouser: migrate to scoped plugin-sdk imports * Chore: remove accidental .DS_Store artifact * chore(docs): add plugins refactor changelog entry * feat(ios): add Live Activity connection status + stale cleanup (#33591) * feat(ios): add live activity connection status and cleanup Add lock-screen/Dynamic Island connection health states and prune duplicate/stale activities before reuse. This intentionally excludes AI/title generation and heavier UX rewrites from #27488. Co-authored-by: leepokai <[email protected]> * fix(ios): treat ended live activities as inactive * chore(changelog): add PR reference and author thanks --------- Co-authored-by: leepokai <[email protected]> * fix: kill stuck ACP child processes on startup and harden sessions in discord threads (#33699) * Gateway: resolve agent.wait for chat.send runs * Discord: harden ACP thread binding + listener timeout * ACPX: handle already-exited child wait * Gateway/Discord: address PR review findings * Discord: keep ACP error-state thread bindings on startup * gateway: make agent.wait dedupe bridge event-driven * discord: harden ACP probe classification and cap startup fan-out * discord: add cooperative timeout cancellation * discord: fix startup probe concurrency helper typing * plugin-sdk: avoid Windows root-alias shard timeout * plugin-sdk: keep root alias reflection path non-blocking * discord+gateway: resolve remaining PR review findings * gateway+discord: fix codex review regressions * Discord/Gateway: address Codex review findings * Gateway: keep agent.wait lifecycle active with shared run IDs * Discord: clean up status reactions on aborted runs * fix: add changelog note for ACP/Discord startup hardening (#33699) (thanks @dutifulbob) --------- Co-authored-by: Onur <[email protected]> * fix: relay ACP sessions_spawn parent streaming (#34310) (thanks @vincentkoc) (#34310) Co-authored-by: Onur Solmaz <[email protected]> * fix(telegram): materialize dm draft final to avoid duplicates * docs(changelog): credit @Brotherinlaw-13 for #34318 * fix: prevent nodes media base64 context bloat (#34332) * fix: preserve raw media invoke for HTTP tool clients (#34365) * fix(slack): route system events to bound agent sessions (#34045) * fix(slack): route system events via binding-aware session keys * fix(slack): pass sender to system event session resolver * fix(slack): include sender context for interaction session routing * fix(slack): include modal submitter in session routing * test(slack): cover binding-aware system event routing * test(slack): update interaction session key assertions * test(slack): assert reaction session routing carries sender * docs(changelog): note slack system event routing fix * Update CHANGELOG.md * Delete changelog/fragments directory * fix(memory): serialize local embedding initialization to avoid duplicate model loads (#15639) Merged via squash. Prepared head SHA: a085fc21a8ba7163fffdb5de640dd4dc1ff5a88e Co-authored-by: SubtleSpark <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(model): propagate custom provider headers to model objects (#27490) Merged via squash. Prepared head SHA: e4183b398fc7eb4c18b2b691cb0dd882ec993608 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(daemon): handle systemctl is-enabled exit 4 (not-found) on Ubuntu (#33634) Merged via squash. Prepared head SHA: 67dffc3ee239cd7b813cb200c3dd5475d9e203a6 Co-authored-by: Yuandiaodiaodiao <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(node-host): sync rawCommand with hardened argv after executable path pinning (#33137) Merged via squash. Prepared head SHA: a7987905f7ad6cf5fee286ffa81ceaad8297174f Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * Agents: add generic poll-vote action support * fix(ollama): pass provider headers to Ollama stream function (#24285) createOllamaStreamFn() only accepted baseUrl, ignoring custom headers configured in models.providers.<provider>.headers. This caused 403 errors when Ollama endpoints are behind reverse proxies that require auth headers (e.g. X-OLLAMA-KEY via HAProxy). Add optional defaultHeaders parameter to createOllamaStreamFn() and merge them into every fetch request. Provider headers from config are now passed through at the call site in the embedded runner. Fixes #24285 * test(ollama): add default header precedence coverage * chore(changelog): add PR entry openclaw#24337 thanks @echoVic * Outbound: allow text-only plugin adapters * Outbound: avoid empty multi-media fallback sends * chore(changelog): align outbound adapter entry openclaw#32788 thanks @liuxiaopai-ai * fix(outbound): fail media-only text-only adapter fallback * chore(changelog): clarify outbound media-only fallback openclaw#32788 thanks @liuxiaopai-ai * fix(review): enforce behavioral sweep validation * Fix gateway restart false timeouts on Debian/systemd (#34874) * daemon(systemd): target sudo caller user scope * test(systemd): cover sudo user scope commands * infra(ports): fall back to ss when lsof missing * test(ports): verify ss fallback listener detection * cli(gateway): use probe fallback for restart health * test(gateway): cover restart-health probe fallback * Compaction/Safeguard: require structured summary headings (#25555) Merged via squash. Prepared head SHA: 0b1df34806a7b788261290be55760fd89220de53 Co-authored-by: rodrigouroz <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * Fix Linux daemon install checks when systemd user bus env is missing (#34884) * daemon(systemd): fall back to machine user scope when user bus is missing * test(systemd): cover machine scope fallback for user-bus errors * test(systemd): reset execFile mock state across cases * test(systemd): make machine-user fallback assertion portable * fix(daemon): keep root sudo path on direct user scope * test(systemd): cover sudo root user-scope behavior * ci: use resolvable bun version in setup-node-env * agents: preserve totalTokens on request failure instead of using contextWindow (#34275) Merged via squash. Prepared head SHA: f9d111d0a79a07815d476356e98a28df3a0000ba Co-authored-by: RealKai42 <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix: align AGENTS.md template section names with post-compaction extraction (#25029) (#25098) Merged via squash. Prepared head SHA: 8cd6cc8049aab5a94d8a9d5fb08f2e792c4ac5fd Co-authored-by: echoVic <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * Changelog: add daemon systemd user-bus fallback entry (#34884) * Changelog: add gateway restart health entry (#34874) * fix: finalize spanish locale support * fix: add spanish locale support (#35038) (thanks @DaoPromociones) * fix(deps): patch hono transitive audit vulnerabilities * fix(security): avoid prototype-chain account path checks (#34982) Merged via squash. Prepared head SHA: f89cc6a649959997fe1dec1e1c1bff9a61b2de98 Co-authored-by: HOYALIM <[email protected]> Co-authored-by: dvrshil <[email protected]> Reviewed-by: @dvrshil * fix(deps): bump tar to 7.5.10 * docs(changelog): document dependency security fixes * fix: restore auto-reply system events timeline (#34794) (thanks @anisoptera) (#34794) Co-authored-by: Ayaan Zaidi <[email protected]> * fix(feishu): comprehensive reply mechanism — outbound replyToId forwarding + topic-aware reply targeting (#33789) * fix(feishu): comprehensive reply mechanism fix — outbound replyToId forwarding + topic-aware reply targeting - Forward replyToId from ChannelOutboundContext through sendText/sendMedia to sendMessageFeishu/sendMarkdownCardFeishu/sendMediaFeishu, enabling reply-to-message via the message tool. - Fix group reply targeting: use ctx.messageId (triggering message) in normal groups to prevent silent topic thread creation (#32980). Preserve ctx.rootId targeting for topic-mode groups (group_topic/group_topic_sender) and groups with explicit replyInThread config. - Add regression tests for both fixes. Fixes #32980 Fixes #32958 Related #19784 * fix: normalize Feishu delivery.to before comparing with messaging tool targets - Add normalizeDeliveryTarget helper to strip user:/chat: prefixes for Feishu - Apply normalization in matchesMessagingToolDeliveryTarget before comparison - This ensures cron duplicate suppression works when session uses prefixed targets (user:ou_xxx) but messaging tool extract uses normalized bare IDs (ou_xxx) Fixes review comment on PR #32755 (cherry picked from commit fc20106f16ccc88a5f02e58922bb7b7999fe9dcd) * fix(feishu): catch thrown SDK errors for withdrawn reply targets The Feishu Lark SDK can throw exceptions (SDK errors with .code or AxiosErrors with .response.data.code) for withdrawn/deleted reply targets, in addition to returning error codes in the response object. Wrap reply calls in sendMessageFeishu and sendCardFeishu with try-catch to handle thrown withdrawn/not-found errors (230011, 231003) and fall back to client.im.message.create, matching the existing response-level fallback behavior. Also extract sendFallbackDirect helper to deduplicate the direct-send fallback block across both functions. Closes #33496 (cherry picked from commit ad0901aec103a2c52f186686cfaf5f8ba54b4a48) * feishu: forward outbound reply target context (cherry picked from commit c129a691fcf552a1cebe1e8a22ea8611ffc3b377) * feishu extension: tighten reply target fallback semantics (cherry picked from commit f85ec610f267020b66713c09e648ec004b2e26f1) * fix(feishu): align synthesized fallback typing and changelog attribution * test(feishu): cover group_topic_sender reply targeting --------- Co-authored-by: Xu Zimo <[email protected]> Co-authored-by: Munem Hashmi <[email protected]> Co-authored-by: bmendonca3 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(feishu): use msg_type media for mp4 video (fixes #33674) (#33720) * fix(feishu): use msg_type media for mp4 video (fixes #33674) * Feishu: harden streaming merge semantics and final reply dedupe Use explicit streaming update semantics in the Feishu reply dispatcher: treat onPartialReply payloads as snapshot updates and block fallback payloads as delta chunks, then merge final text with the shared overlap-aware mergeStreamingText helper before closing the stream. Prevent duplicate final text delivery within the same dispatch cycle, and add regression tests covering overlap snapshot merge, duplicate final suppression, and block-as-delta behavior to guard against repeated/truncated output. * fix(feishu): prefer message.reply for streaming cards in topic threads * fix: reduce Feishu streaming card print_step to avoid duplicate rendering Fixes openclaw/openclaw#33751 * Feishu: preserve media sends on duplicate finals and add media synthesis changelog * Feishu: only dedupe exact duplicate final replies * Feishu: use scoped plugin-sdk import in streaming-card tests --------- Co-authored-by: 倪汉杰0668001185 <[email protected]> Co-authored-by: zhengquanliu <[email protected]> Co-authored-by: nick <[email protected]> Co-authored-by: linhey <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(agents): bypass pendingDescendantRuns guard for cron announce delivery (#35185) * fix(agents): bypass pendingDescendantRuns guard for cron announce delivery Standalone cron job completions were blocked from direct channel delivery when the cron run had spawned subagents that were still registered as pending. The pendingDescendantRuns guard exists for live orchestration coordination and should not apply to fire-and-forget cron announce sends. Thread the announceType through the delivery chain and skip both the child-descendant and requester-descendant pending-run guards when the announce originates from a cron job. Closes #34966 * fix: ensure outbound session entry for cron announce with named agents (#32432) Named agents may not have a session entry for their delivery target, causing the announce flow to silently fail (delivered=false, no error). Two fixes: 1. Call ensureOutboundSessionEntry when resolving the cron announce session key so downstream delivery can find channel metadata. 2. Fall back to direct outbound delivery when announce delivery fails to ensure cron output reaches the target channel. Closes #32432 Co-Authored-By: Claude Opus 4.6 <[email protected]> * fix: guard announce direct-delivery fallback against suppression leaks (#32432) The `!delivered` fallback condition was too broad — it caught intentional suppressions (active subagents, interim messages, SILENT_REPLY_TOKEN) in addition to actual announce delivery failures. Add an `announceDeliveryWasAttempted` flag so the direct-delivery fallback only fires when `runSubagentAnnounceFlow` was actually called and failed. Also remove the redundant `if (route)` guard in `resolveCronAnnounceSessionKey` since `resolved` being truthy guarantees `route` is non-null. Co-Authored-By: Claude Opus 4.6 <[email protected]> * fix(cron): harden announce synthesis follow-ups --------- Co-authored-by: scoootscooob <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * Feishu: harden streaming merge semantics and final reply dedupe (#33245) * Feishu: close duplicate final gap and cover routing precedence * Feishu: resolve reviewer duplicate-final and routing feedback * Feishu: tighten streaming send-mode option typing * Feishu: fix reverse-overlap streaming merge ordering * Feishu: align streaming final dedupe test expectation * Feishu: allow distinct streaming finals while deduping repeats --------- Co-authored-by: Tak Hoffman <[email protected]> * fix: cron backup should preserve pre-edit snapshot (#35195) (#35234) * fix(cron): avoid overwriting .bak during normalization Fixes openclaw/openclaw#35195 * test(cron): preserve pre-edit bak snapshot in normalization path --------- Co-authored-by: 0xsline <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(cron): stabilize restart catch-up replay semantics (#35351) * Cron: stabilize restart catch-up replay semantics * Cron: respect backoff in startup missed-run replay * cron: narrow startup replay backoff guard (#35391) * cron: unify stale-run recovery and preserve manual-run every anchors (#35363) * cron: unify stale-run recovery and preserve manual every anchors * cron: address unresolved review threads on recovery paths * cron: remove duplicate timestamp helper after rebase * refactor(telegram): remove unused webhook callback helper (#27816) * fix(pr): make review claim step required * fix(skills): deduplicate slash commands by skillName across all interfaces Move skill-command deduplication by skillName from the Discord-only `dedupeSkillCommandsForDiscord` into `listSkillCommandsForAgents` so every interface (TUI, Slack, text) consistently sees a clean command list without platform-specific workarounds. When multiple agents share a skill with the same name the old code emitted `github` + `github_2` and relied on Discord to collapse them. Now `listSkillCommandsForAgents` returns only the first registration per skillName, and the Discord-specific wrapper is removed. Co-Authored-By: Claude Sonnet 4.6 <[email protected]> * style: fix formatting in skill-commands.test.ts and provider.ts Co-Authored-By: Claude Sonnet 4.6 <[email protected]> * style(skills): align formatting cleanup for dedupe changes * chore(changelog): add dedupe note openclaw#27521 thanks @shivama205 * fix(agents): detect Venice provider proxying xAI/Grok models for schema cleaning (#35355) Merged via squash. Prepared head SHA: 8bfdec257bb6a6025cb69a0a213a433da32b15db Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(agents): decode HTML entities in xAI/Grok tool call arguments (#35276) Merged via squash. Prepared head SHA: c4445d2938898ded9c046614f9315dbda65ec573 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(agents): guard promoteThinkingTagsToBlocks against malformed content entries (#35143) Merged via squash. Prepared head SHA: 3971122f5fd27c66c8c9c5ce783f00e113b1f47b Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(web-ui): render Accounts schema node properly (#35380) Co-authored-by: stakeswky <[email protected]> Co-authored-by: liuxiaopai-ai <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(agents): guard context pruning against malformed thinking blocks (#35146) Merged via squash. Prepared head SHA: a196a565b1b8e806ffbf85172bcf1128796b45a2 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * fix(gateway): prevent internal route leakage in chat.send Synthesis of routing fixes from #35321, #34635, and #35356 for internal-client reply safety. - Require explicit `deliver: true` before inheriting any external delivery route. - Keep webchat/TUI/UI-origin traffic on internal routing by default. - Allow configured-main session inheritance only for non-Webchat/UI clients, and honor `session.mainKey`. - Add regression tests for UI no-inherit, configured-main CLI inherit, and deliver-flag behavior. Co-authored-by: alexyyyander <[email protected]> Co-authored-by: Octane0411 <[email protected]> Co-authored-by: Linux2010 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(gateway): pass actual version to Control UI client instead of dev (#35230) * fix(gateway): pass actual version to Control UI client instead of "dev" The GatewayClient, CLI WS client, and browser Control UI all sent "dev" as their clientVersion during handshake, making it impossible to distinguish builds in gateway logs and health snapshots. - GatewayClient and CLI WS client now use the resolved VERSION constant - Control UI reads serverVersion from the bootstrap endpoint and forwards it when connecting - Bootstrap contract extended with serverVersion field Closes #35209 * Gateway: fix control-ui version version-reporting consistency * Control UI: guard deferred bootstrap connect after disconnect * fix(ui): accept same-origin http and relative gateway URLs for client version --------- Co-authored-by: Tak Hoffman <[email protected]> * chore(pr): enforce changelog placement and reduce merge sync churn * TTS: add baseUrl support to OpenAI TTS config (#34321) Merged via squash. Prepared head SHA: e9a10cf81d2021cf81091dfa81e13ffdbb6a540a Co-authored-by: RealKai42 <[email protected]> Co-authored-by: shakkernerd <[email protected]> Reviewed-by: @shakkernerd * ACP: add persistent Discord channel and Telegram topic bindings (#34873) * docs: add ACP persistent binding experiment plan * docs: align ACP persistent binding spec to channel-local config * docs: scope Telegram ACP bindings to forum topics only * docs: lock bound /new and /reset behavior to in-place ACP reset * ACP: add persistent discord/telegram conversation bindings * ACP: fix persistent binding reuse and discord thread parent context * docs: document channel-specific persistent ACP bindings * ACP: split persistent bindings and share conversation id helpers * ACP: defer configured binding init until preflight passes * ACP: fix discord thread parent fallback and explicit disable inheritance * ACP: keep bound /new and /reset in-place * ACP: honor configured bindings in native command flows * ACP: avoid configured fallback after runtime bind failure * docs: refine ACP bindings experiment config examples * acp: cut over to typed top-level persistent bindings * ACP bindings: harden reset recovery and native command auth * Docs: add ACP bound command auth proposal * Tests: normalize i18n registry zh-CN assertion encoding * ACP bindings: address review findings for reset and fallback routing * ACP reset: gate hooks on success and preserve /new arguments * ACP bindings: fix auth and binding-priority review findings * Telegram ACP: gate ensure on auth and accepted messages * ACP bindings: fix session-key precedence and unavailable handling * ACP reset/native commands: honor fallback targets and abort on bootstrap failure * Config schema: validate ACP binding channel and Telegram topic IDs * Discord ACP: apply configured DM bindings to native commands * ACP reset tails: dispatch through ACP after command handling * ACP tails/native reset auth: fix target dispatch and restore full auth * ACP reset detection: fallback to active ACP keys for DM contexts * Tests: type runTurn mock input in ACP dispatch test * ACP: dedup binding route bootstrap and reset target resolution * reply: align ACP reset hooks with bound session key * docs: replace personal discord ids with placeholders * fix: add changelog entry for ACP persistent bindings (#34873) (thanks @dutifulbob) --------- Co-authored-by: Onur <[email protected]> * docs(telegram): recommend allowlist for single-user DM policy (#34841) * docs(telegram): recommend allowlist for single-user bots * docs(telegram): condense single-user allowlist note --------- Co-authored-by: echoVic <[email protected]> * fix(feishu): check response.ok before calling response.json() in streaming card (#35628) Merged via squash. Prepared head SHA: 62c3fec80d97cea9be344c0bef5358a0a5dc5560 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * Mattermost: honor onmessage mention override and add gating diagnostics tests (#27160) Merged via squash. Prepared head SHA: 6cefb1d5bf3d6dfcec36c1cee3f9ea887f10c890 Co-authored-by: turian <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Reviewed-by: @mukhtharcm * fix(subagents): strip leaked [[reply_to]] tags from completion announces (#34503) * fix(subagents): strip reply tags from completion delivery text * test(subagents): cover reply-tag stripping in cron completion sends * changelog: note iMessage reply-tag stripping in completion announces * Update CHANGELOG.md * Apply suggestion from @greptile-apps[bot] Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> --------- Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix(cron): restore direct fallback after announce failure in best-effort mode (openclaw#36177) Verified: - pnpm build - pnpm check (fails on pre-existing origin/main lint debt in extensions/mattermost imports) - pnpm test:macmini Co-authored-by: Tak Hoffman <[email protected]> * test(cron): add cross-channel announce fallback regression coverage (openclaw#36197) Verified: - pnpm install --frozen-lockfile - pnpm build - pnpm check (fails on pre-existing origin/main lint debt in extensions/mattermost imports) - pnpm test:macmini Co-authored-by: Tak Hoffman <[email protected]> * feat(mattermost): add interactive buttons support (#19957) Merged via squash. Prepared head SHA: 8a25e608729d0b9fd07bb0ee4219d199d9796dbe Co-authored-by: tonydehnke <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Reviewed-by: @mukhtharcm * fix(browser): remove deprecated --disable-blink-features=AutomationControlled flag - Removes OpenClaw's default `--disable-blink-features=AutomationControlled` Chrome launch switch to avoid unsupported-flag warnings in newer Chrome (#35721). - Preserves compatibility for older Chrome via `browser.extraArgs` override behavior (source analysis: #35770, #35728, #35727, #35885). - Synthesis attribution: thanks @Sid-Qin, @kevinWangSheng, @ningding97, @Naylenv, @clawbie. Source PR refs: #35734, #35770, #35728, #35727, #35885 Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: kevinWangSheng <[email protected]> Co-authored-by: ningding97 <[email protected]> Co-authored-by: Naylenv <[email protected]> Co-authored-by: clawbie <[email protected]> Co-authored-by: Takhoffman <[email protected]> * fix(feishu): add HTTP timeout to prevent per-chat queue deadlocks (#36430) When the Feishu API hangs or responds slowly, the sendChain never settles, causing the per-chat queue to remain in a processing state forever and blocking all subsequent messages in that thread. This adds a 30-second default timeout to all Feishu HTTP requests by providing a timeout-aware httpInstance to the Lark SDK client. Closes #36412 Co-authored-by: Ayane <[email protected]> * fix(feishu): use probed botName for mention checks (#36391) * Feishu: honor bot mentions by ID despite aliases (Fixes #36317) (#36333) * Mattermost: switch plugin-sdk imports to scoped subpaths (openclaw#36480) Verified: - pnpm build - pnpm check - pnpm test:macmini Co-authored-by: Takhoffman <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(feishu): accept groupPolicy "allowall" as alias for "open" (#36358) * fix(feishu): accept groupPolicy "allowall" as alias for "open" When users configure groupPolicy: "allowall" in Feishu channel config, the Zod schema rejects the value and the runtime policy check falls through to the allowlist path. With an empty allowFrom array, all group messages are silently dropped despite the intended "allow all" semantics. Accept "allowall" at the schema level (transform to "open") and add a runtime guard in isFeishuGroupAllowed so the value is handled even if it bypasses schema validation. Closes #36312 Made-with: Cursor * Feishu: tighten allowall alias handling and coverage --------- Co-authored-by: Tak Hoffman <[email protected]> * synthesis: fix Feishu group mention slash parsing ## Summary\n\nFeishu group slash command parsing is fixed for mentions and command probes across authorization paths.\n\nThis includes:\n- Normalizing bot mention text in group context for reliable slash detection in message parsing.\n- Adding command-probe normalization for group slash invocations.\n\nCo-authored-by: Sid Qin <[email protected]>\nCo-authored-by: Tak Hoffman <[email protected]> * Feishu: normalize group slash command probing - Feishu/group slash command detection: normalize group mention wrappers before command-authorization probing so mention-prefixed commands are recognized in group routing.\n- Source PR: #36011\n- Contributor: @liuxiaopai-ai\n\nCo-authored-by: Tak Hoffman <[email protected]>\nCo-authored-by: liuxiaopai-ai <[email protected]> * add prependSystemContext and appendSystemContext to before_prompt_build (fixes #35131) (#35177) Merged via squash. Prepared head SHA: d9a2869ad69db9449336a2e2846bd9de0e647ac6 Co-authored-by: maweibin <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(feishu): avoid media regressions from global HTTP timeout (#36500) * fix(feishu): avoid media regressions from global http timeout * fix(feishu): source HTTP timeout from config * fix(feishu): apply media timeout override to image uploads * fix(feishu): invalidate cached client when timeout changes * fix(feishu): clamp timeout values and cover image download * Gateway: add SecretRef support for gateway.auth.token with auth-mode guardrails (#35094) * fix(embedded): classify model_context_window_exceeded as context overflow, trigger compaction (#35934) Merged via squash. Prepared head SHA: 20fa77289c80b2807a6779a3df70440242bc18ca Co-authored-by: RealKai42 <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(agents): skip compaction API call when session has no real messages (#36451) Merged via squash. Prepared head SHA: 52dd6317895c7bd10855d2bd7dbbfc2f5279b68e Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(ui): catch marked.js parse errors to prevent Control UI crash (#36445) - Prevent Control UI session render crashes when `marked.parse()` encounters pathological recursive markdown by safely falling back to escaped `<pre>` output. - Tighten markdown fallback regression coverage and keep changelog attribution in sync for this crash-hardening path. Co-authored-by: Bin Deng <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(session): archive old transcript on daily/scheduled reset to prevent orphaned files (#35493) Merged via squash. Prepared head SHA: 0d95549d752adecfc0b08d5cd55a8b8c75e264fe Co-authored-by: byungsker <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(agents): set preserveSignatures to isAnthropic in resolveTranscriptPolicy (#32813) Merged via squash. Prepared head SHA: f522d21ca59a42abac554435a0aa646f6a34698d Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix: avoid false global rate-limit classification from generic cooldown text (#32972) Merged via squash. Prepared head SHA: 813c16f5afce415da130a917d9ce9f968912b477 Co-authored-by: stakeswky <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * refactor(agents): share failover HTTP status classification (#36615) * fix(agents): classify transient failover statuses consistently * fix(agents): preserve legacy failover status mapping * fix(failover): narrow service-unavailable to require overload indicator (#32828) (#36646) Merged via squash. Prepared head SHA: 46fb4306127972d7635f371fd9029fbb9baff236 Co-authored-by: jnMetaCode <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * Compaction/Safeguard: add summary quality audit retries (#25556) Merged via squash. Prepared head SHA: be473efd1635616ebbae6e649d542ed50b4a827f Co-authored-by: rodrigouroz <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * test(agents): add provider-backed failover regressions (#36735) * test(agents): add provider-backed failover fixtures * test(agents): cover more provider error docs * test(agents): tighten provider doc fixtures * Docs: add Slack typing reaction fallback * Docs: update gateway config reference for Slack and TTS * Docs: clarify OpenAI-compatible TTS endpoints * Docs: document Control UI locale support * Docs: cover heartbeat, cron, and plugin route updates * fix(ui): bump dompurify to 3.3.2 (#36781) * UI: bump dompurify to 3.3.2 * Deps: refresh dompurify lockfile * UI: hoist lifecycle connect test mocks (#36788) * fix(agents): classify insufficient_quota 400s as billing (#36783) * feat: append UTC time alongside local time in shared Current time lines (#32423) Merged via squash. Prepared head SHA: 9e8ec13933b5317e7cff3f0bc048de515826c31a Co-authored-by: jriff <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * fix(auth): grant senderIsOwner for internal channels with operator.admin scope (openclaw#35704) Verified: - pnpm install --frozen-lockfile - pnpm build - pnpm check - pnpm test:macmini Co-authored-by: Naylenv <[email protected]> Co-authored-by: Octane0411 <[email protected]> Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(config): prevent RangeError in merged schema cache key generation Fix merged schema cache key generation for high-cardinality plugin/channel metadata by hashing incrementally instead of serializing one large aggregate string. Includes changelog entry for the user-visible regression fix. Co-authored-by: Tak Hoffman <[email protected]> Co-authored-by: Bill <[email protected]> * fix(slack): propagate mediaLocalRoots through Slack send path Restore Slack local file upload parity with CVE-era local media allowlist enforcement by threading `mediaLocalRoots` through the Slack send call chain. - pass `ctx.mediaLocalRoots` from Slack channel action adapter into `handleSlackAction` - add and forward `mediaLocalRoots` in Slack action context/send path - pass `mediaLocalRoots` into `sendMessageSlack` for upload allowlist enforcement - add changelog entry with attribution for this behavior fix Co-authored-by: 2233admin <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(slack): preserve dedupe while recovering dropped app_mention (#34937) This PR fixes Slack mention loss without reintroducing duplicate dispatches. - Preserve seen-message dedupe at ingress to prevent duplicate processing. - Allow a one-time app_mention retry only when the paired message event was previously dropped before dispatch. - Add targeted race tests for both recovery and duplicate-prevention paths. Co-authored-by: littleben <[email protected]> Co-authored-by: OpenClaw Agent <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * README: add algal to contributors list (#2046) * fix: decouple Discord inbound worker timeout from listener timeout (#36602) (thanks @dutifulbob) (#36602) Co-authored-by: Onur Solmaz <[email protected]> * plugins: enforce prompt hook policy with runtime validation (#36567) Merged via squash. Prepared head SHA: 6b9d883b6ae33628235fb02ce39c0d0f46a065bb Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(memory): avoid destructive qmd collection rebinds * Harden Telegram poll gating and schema consistency (#36547) Merged via squash. Prepared head SHA: f77824419e3d166f727474a9953a063a2b4547f2 Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(browser): close tracked tabs on session cleanup (#36666) * Diffs: restore system prompt guidance (#36904) Merged via squash. Prepared head SHA: 1b3be3c87957c068473d5c86b9efba4a1a8503f2 Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * fix(routing): avoid full binding rescans in resolveAgentRoute (#36915) * fix(gateway): honor insecure ws override for remote hostnames * fix(llm-task): load runEmbeddedPiAgent from dist/extensionAPI in installs * fix(auth): harden openai-codex oauth login path * feat(telegram/acp): Topic Binding, Pin Binding Message, Fix Spawn Param Parsing (#36683) * fix(acp): normalize unicode flags and Telegram topic binding * feat(telegram/acp): restore topic-bound ACP and session bindings * fix(acpx): clarify permission-denied guidance * feat(telegram/acp): pin spawn bind notice in topics * docs(telegram): document ACP topic thread binding behavior * refactor(reply): share Telegram conversation-id resolver * fix(telegram/acp): preserve bound session routing semantics * fix(telegram): respect binding persistence and expiry reporting * refactor(telegram): simplify binding lifecycle persistence * fix(telegram): bind acp spawns in direct messages * fix: document telegram ACP topic binding changelog (#36683) (thanks @huntharo) --------- Co-authored-by: Onur <[email protected]> * fix(gateway): preserve streamed prefixes across tool boundaries * fix(tui): prevent stale model indicator after /model * Memory: handle SecretRef keys in doctor embeddings (#36835) Merged via squash. Prepared head SHA: c1a3d0caae60115d886e8bfc9983c9533c773f04 Co-authored-by: joshavant <[email protected]> Co-authored-by: joshavant <[email protected]> Reviewed-by: @joshavant * fix(openai-codex): request required oauth api scopes (#24720) * fix(memory-flush): ban timestamped variant files in default flush prompt (#34951) Merged via squash. Prepared head SHA: efadda4988b460e6da07be72994d4951d64239d0 Co-authored-by: zerone0x <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(tui): render final event error when assistant output is empty (#14687) * feat(agents): flush reply pipeline before compaction wait (#35489) Merged via squash. Prepared head SHA: 7dbbcc510b74b0e8d35eb750d24575e34b5d769a Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: jalehman <[email protected]> Reviewed-by: @jalehman * fix(secrets): harden api key normalization for ByteString headers * fix(slack): remove double mrkdwn conversion in native streaming path Remove redundant text normalization from Slack native streaming markdown_text flow so Markdown formatting is preserved. Synthesis context: overlaps reviewed from #34931, #34759, #34716, #34682, #34814. Co-authored-by: littleben <[email protected]> Co-authored-by: dunamismax <[email protected]> Co-authored-by: Octane <[email protected]> Co-authored-by: Mitsuyuki Osabe <[email protected]> Co-authored-by: Kai <[email protected]> Co-authored-by: OpenClaw Agent <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> * fix(kimi-coding): normalize anthropic tool payload format * fix(slack): thread channel ID through inbound context for reactions (#34831) Slack reaction/thread context routing fixes via canonical synthesis of #34831. Co-authored-by: Tak <[email protected]> * fix(heartbeat): pin HEARTBEAT.md reads to workspace path * fix(subagents): recover announce cleanup after kill/complete race * feat(hooks): emit compaction lifecycle hooks (#16788) * fix(auth): harden openai-codex oauth refresh fallback * fix(subagents): announce delivery with descendant gating, frozen result refresh, and cron retry (#35080) Thanks @tyler6204 * fix(agents): avoid synthetic tool-result writes on idle-timeout cleanup * fix(agent): harden undici stream timeouts for long openai-completions runs * fix(slack): record app_mention retry key before dedupe check (#37033) - Prime app_mention retry allowance before dedupe so near-simultaneous message/app_mention races do not drop valid mentions. - Prevent duplicate dispatch when app_mention wins the race and message prepare later succeeds. - Prune dispatched mention keys and add regression coverage for both dropped and successful in-flight message outcomes. Co-authored-by: Tak Hoffman <[email protected]> * fix(agents): honor explicit rate-limit cooldown probes in fallback runs * fix(agents): allow configured ollama endpoints without dummy api keys * fix(memory): recover qmd updates from duplicate document constraints * Doctor: warn on implicit heartbeat directPolicy (#36789) * Changelog: note heartbeat directPolicy doctor warning * Tests: cover heartbeat directPolicy doctor warning * Doctor: warn on implicit heartbeat directPolicy * Tests: cover per-agent heartbeat directPolicy warning * Update CHANGELOG.md * Plugins: clarify registerHttpHandler migration errors (#36794) * Changelog: note plugin HTTP route migration diagnostics * Tests: cover registerHttpHandler migration diagnostics * Plugins: clarify registerHttpHandler migration errors * Tests: cover registerHttpHandler diagnostic edge cases * Plugins: tighten registerHttpHandler migration hint * fix(memory): repair qmd collection name conflicts during ensure * fix(memory): handle qmd search results without docid * Plugins: avoid false integrity drift prompts on unpinned updates (#37179) * Plugins: skip drift prompts for unpinned updates * Plugins: cover unpinned integrity update behavior * Changelog: add #37179 release note * Delete changelog/fragments directory * Update CHANGELOG.md * fix(whatsapp): remove implicit [openclaw] self-chat prefix * fix: remove config.schema from agent gateway tool (#7382) Merged via squash. Prepared head SHA: f34a7780690a941936b31899e2d096b8a07f4afc Co-authored-by: kakuteki <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * feat(openai): add gpt-5.4 support for API and Codex OAuth (#36590) * feat(openai): add gpt-5.4 support and priority processing * feat(openai-codex): add gpt-5.4 oauth support * fix(openai): preserve provider overrides in gpt-5.4 fallback * fix(openai-codex): keep xhigh for gpt-5.4 default * fix(models): preserve configured overrides in list output * fix(models): close gpt-5.4 integration gaps * fix(openai): scope service tier to public api * fix(openai): complete prep followups for gpt-5.4 support (#36590) (thanks @dorukardahan) --------- Co-authored-by: Tyler Yust <[email protected]> * fix(tui): preserve credential-like tokens in render sanitization * CLI: make read-only SecretRef status flows degrade safely (#37023) * CLI: add read-only SecretRef inspection * CLI: fix read-only SecretRef status regressions * CLI: preserve read-only SecretRef status fallbacks * Docs: document read-only channel inspection hook * CLI: preserve audit coverage for read-only SecretRefs * CLI: fix read-only status account selection * CLI: fix targeted gateway fallback analysis * CLI: fix Slack HTTP read-only inspection * CLI: align audit credential status checks * CLI: restore Telegram read-only fallback semantics * chore(changelog): update for #37023 Signed-off-by: joshavant <[email protected]> * fix(agents): disable usage streaming chunks on non-native openai-completions * feat(nano-banana-pro): add --aspect-ratio flag to generate_image.py (#28159) * feat(nano-banana-pro): add --aspect-ratio flag to generate_image.py * Nano Banana: allow all supported aspect ratios * Docs: expand nano banana aspect ratio options --------- Co-authored-by: Vincent Koc <[email protected]> * fix(gateway): support image_url in OpenAI chat completions (#34068) * fix(gateway): parse image_url in openai chat completions * test(gateway): cover openai chat completions image_url flows * docs(changelog): note openai image_url chat completions fix (#17685) * fix(gateway): harden openai image_url parsing and limits * test(gateway): add openai image_url regression coverage * docs(changelog): expand #17685 openai chat completions note * Gateway: make OpenAI image_url URL fetch opt-in and configurable * Diagnostics: redact image base64 payload data in trace logs * Changelog: note OpenAI image_url hardening follow-ups * Gateway: enforce OpenAI image_url total budget incrementally * Gateway: scope OpenAI image_url extraction to the active turn * Update CHANGELOG.md * fix(agents): avoid xAI web_search tool-name collisions * fix: clear Telegram DM draft after materialize (#36746) (thanks @joelnishanth) * Fix Control UI duplicate iMessage replies for internal webchat turns (#36151) * Auto-reply: avoid routing external replies from internal webchat turns * Auto-reply tests: cover internal webchat non-routing with external origin metadata * Changelog: add Control UI iMessage duplicate-reply fix note * Auto-reply context: track explicit deliver routes * Gateway chat: mark explicit external deliver routes in context * Auto-reply: preserve explicit deliver routes for internal webchat turns * Auto-reply tests: cover explicit deliver routes from internal webchat turns * Gateway chat tests: assert explicit deliver route context tagging * fix: enforce 600 perms for cron store and run logs (#36078) * fix: enforce secure permissions for cron store and run logs * fix(cron): enforce dir perms and gate posix tests on windows * Cron store tests: cover existing directory permission hardening * Cron run-log tests: cover existing directory permission hardening * Changelog: note cron file permission hardening --------- Co-authored-by: linhey <[email protected]> Co-authored-by: Vincent Koc <[email protected]> * fix(tui): accept canonical session-key aliases in chat event routing * Gateway: normalize OpenAI stream chunk text * Gateway: coerce chat deliverable route boolean * fix(web_search): align brave language codes with API * Respect source channel for agent event surfacing (#36030) * fix(session): prefer webchat routes for direct ui turns (#37135) * Gateway: discriminate input sources * Cron: migrate legacy provider delivery hints * Cron: stabilize runs-one-shot migration tests * fix(memory): retry mcporter after Windows EINVAL spawn * fix(onboarding): guard daemon status probe on headless linux * Gateway: add path-scoped config schema lookup (#37266) Merged via squash. Prepared head SHA: 0c4d187f6fb66f2799d4047585d6368e433c883a Co-authored-by: gumadeiras <[email protected]> Co-authored-by: gumadeiras <[email protected]> Reviewed-by: @gumadeiras * docs(changelog): add pr entry * fix(ci): restore protocol and schema checks (#37470) * Fix failover for zhipuai 1310 Weekly/Monthly Limit Exhausted (#33813) Merged via squash. Prepared head SHA: 3dc441e58de48913720cf7b6137fa761758d8344 Co-authored-by: zhouhe-xydt <[email protected]> Co-authored-by: altaywtf <[email protected]> Reviewed-by: @altaywtf * fix(openai-codex-oauth): stop mutating authorize url scopes * Update CHANGELOG.md * fix(auth): remove bogus codex oauth responses probe * docs(changelog): fold codex oauth fix notes * docs(changelog): add codex oauth pr reference (#37558) * fix(security:PLA-697): block local secret scratch files from commits * feat: add dev EKS deployment for openclaw SRE container * fix: stabilize slack socket mode in dev eks runtime * fix(eks): reuse monitoring incident auth for openclaw-sre * feat(deploy): add grafana env guards and prod deploy wrapper * fix(security): redact tool outputs and enforce secret-safe runtime defaults * feat(sre): harden heartbeat routing and enrich triage signals * docs: add SRE hybrid intelligence design Three-layer architecture to improve bot reasoning quality: - Layer 1: Service knowledge (auto-discovery + overlays + incident memory) - Layer 2: Multi-stage reasoning chain (triage → hypothesize → causal chain → action plan → cross-review) - Layer 3: Incident learning loop (structured cards, overlay suggestions, feedback signals) Co-Authored-By: Claude Opus 4.6 <[email protected]> * feat(sandbox): add boundary cli to common runtime image * docs(sre): finalize hybrid intelligence design v19 after 18 Codex review rounds Iteratively hardened the design through 18 adversarial Codex review rounds, resolving 60+ findings (15+ CRITICAL, 45+ HIGH). Key fixes include: - Decoupled incident_id (pre-Step11 immutable) from card_id (LLM-derived) - Unified evidence/memory sanitization via shared _strip_instruction_tokens - Fixed rerun interval default (3600s > heartbeat 1800s) to prevent every-cycle re-runs - Added Phase 2 cross-review gap note and dual-column depth table - Overlay suggestion idempotency via deterministic suggestion_key + upsert - Decoupled chain timeout from legacy budget check - Added predicate alignment docs for L3 pre-check vs L3 gate - Adopted dynamic evidence_completeness denominator matching existing code - Added scope note clarifying design vs current implementation Co-Authored-By: Claude Opus 4.6 <[email protected]> * feat(sre): complete hybrid intelligence rollout and bot hardening * fix(sre): harden slack reasoning, dual chain, and deploy auth/runtime * ci: add ecr release pipeline and infra-helm image update trigger * ci: switch workflows to github-hosted runners * ci: allow manual ecr env target and use updater v4 * ci: remove windows jobs from workflow * chore(pla-678): update fallback auth fixture and plan docs * fix(pla-678): preserve control-ui host-header fallback in dev config * refactor(pla-678): remove local sre chart and require infra-helm * refactor(pla-678): remove direct eks deploy scripts * ci(pla-678): remove labeler workflow * ci(pla-678): refresh checks after disabling labeler * build: install boundary CLI in SRE runtime image * fix(slack:PLA-678): keep one thread lane across users * fix(sre-skill): enforce live db-query runbook in dev seed * fix(slack): ignore bot thread history when seeding new sessions * fix(slack): force retry on repeated thread questions * fix(sre-skill): require retry on repeated db requests * docs(sre-skill): make retry policy generic across repeated asks * fix(sre:PLA-678): enforce live linear ticket updates * fix(sre:PLA-678): use [PLATFORM] Backlog linear project * fix(sre:PLA-678): label linked linear tickets as openclaw-sre * fix(sre:PLA-678): auto-label linked linear tickets for auto-pr * test(sre:PLA-678): make tracking label test executable * fix(slack:PLA-678): remove status-final completion banner (#16) * fix(slack:PLA-678): enable direct file attachments from agent replies (#17) * fix(slack:PLA-678): remove status-final completion banner * fix(slack): enable direct file attachments from agent replies * refactor(sre:PLA-678): remove local deploy config mirror (#18) * fix: compact oversized auto-generated pr bodies * fix(sre): add built-in linear incident memory provider * feat(sre): add erpc helper with flo secret query * fix(sre): enforce canonical rpc.morpho.dev erpc endpoint * refactor(sre): remove local deploy config mirror * fix(slack:PLA-678): force DM replies into per-message threads (#19) * fix: compact oversized auto-generated pr bodies * fix(sre): add built-in linear incident memory provider * feat(sre): add erpc helper with flo secret query * fix(sre): enforce canonical rpc.morpho.dev erpc endpoint * refactor(sre): remove local deploy config mirror * fix(slack): force threaded replies for direct messages * feat(cron:PLA-678): use conversation history in self-improve runs (#20) * fix(slack:PLA-678): improve progress update readability (#21) * fix(ci:PLA-678): restore main release pipeline health (#22) * fix(ci:PLA-678): restore main release pipeline health * fix(ci:PLA-678): harden auto-response token fallback * fix(ci:PLA-678): pin setup-bun to published release * test(ci:PLA-678): align slack thread session expectations * fix(sre:PLA-678): harden Slack reply delivery and image release flow (#23) * fix(slack:PLA-678): harden thread-scoped reply delivery * fix(slack:PLA-678): satisfy lint after rebase * ci(release:PLA-678): move image release to cached ecr flow * ci(release:PLA-678): use official docker build actions * fix(slack:PLA-722): preserve progress updates during final-answer gating (#24) * fix(sre:PLA-724): restore ECR runtime image toolchain (#25) * fix(failover): classify HTTP 402 as rate_limit when payload indicates usage limit (#30484) (#36802) * fix(failover): classify HTTP 402 as rate_limit when payload indicates usage limit (#30484) Some providers (notably Anthropic Claude Max plan) surface temporary usage/rate-limit failures as HTTP 402 instead of 429. Before this change, all 402s were unconditionally mapped to 'billing', which produced a misleading 'run out of credits' warning for Max plan users who simply hit their usage window. This follows the same pattern introduced for HTTP 400 in #36783: check the error message for an explicit rate-limit signal before falling back to the default status-code classification. - classifyFailoverReasonFromHttpStatus now returns 'rate_limit' for 402 when isRateLimitErrorMessage matches the payload text - Added regression tests covering both the rate-limit and billing paths on 402 * fix: narrow 402 rate-limit matcher to prevent billing misclassification The original implementation used isRateLimitErrorMessage(), which matches phrases like 'quota exceeded' that legitimately appear in billing errors. This commit replaces it with a narrow, 402-specific matcher that requires BOTH retry language (try again/retry/temporary/cooldown) AND limit terminology (usage limit/rate limit/organization usage). Prevents misclassification of errors like: 'HTTP 402: exceeded quota, please add credits' -> billing (not rate_limit) Added regression test for the ambiguous case. --------- Co-authored-by: Val Alexander <[email protected]> * fix(mattermost): allow reachable interaction callback URLs (#37543) Merged via squash. Prepared head SHA: 4d593731be5a5dcbf3106d596b38acfeb8cf0aa8 Co-authored-by: mukhtharcm <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Reviewed-by: @mukhtharcm * fix(auth): prefer Anthropic API key over stale OAuth profiles * fix(runtime): raise default agent timeout to one hour * feat(runtime): support direct env fallback for agent timeout --------- Signed-off-by: joshavant <[email protected]> Co-authored-by: Gustavo Madeira Santana <[email protected]> Co-authored-by: Mariano <[email protected]> Co-authored-by: leepokai <[email protected]> Co-authored-by: Bob <[email protected]> Co-authored-by: Onur <[email protected]> Co-authored-by: Ayaan Zaidi <[email protected]> Co-authored-by: Vincent Koc <[email protected]> Co-authored-by: huangcj <[email protected]> Co-authored-by: gumadeiras <[email protected]> Co-authored-by: Sid <[email protected]> Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: shakkernerd <[email protected]> Co-authored-by: a <[email protected]> Co-authored-by: echoVic <[email protected]> Co-authored-by: Shakker <[email protected]> Co-authored-by: liuxiaopai-ai <[email protected]> Co-authored-by: Rodrigo Uroz <[email protected]> Co-authored-by: rodrigouroz <[email protected]> Co-authored-by: jalehman <[email protected]> Co-authored-by: Kai <[email protected]> Co-authored-by: RealKai42 <[email protected]> Co-authored-by: 青雲 <[email protected]> Co-authored-by: echoVic <[email protected]> Co-authored-by: Darshil <[email protected]> Co-authored-by: Ho Lim <[email protected]> Co-authored-by: dvrshil <[email protected]> Co-authored-by: Isis Anisoptera <[email protected]> Co-authored-by: Madoka <[email protected]> Co-authored-by: Xu Zimo <[email protected]> Co-authored-by: Munem Hashmi <[email protected]> Co-authored-by: bmendonca3 <[email protected]> Co-authored-by: Tak Hoffman <[email protected]> Co-authored-by: Nhj <[email protected]> Co-authored-by: 倪汉杰0668001185 <[email protected]> Co-authored-by: zhengquanliu <[email protected]> Co-authored-by: nick <[email protected]> Co-authored-by: linhey <[email protected]> Co-authored-by: scoootscooob <[email protected]> Co-authored-by: Claude Opus 4.6 <[email protected]> Co-authored-by: rexl2018 <[email protected]> Co-authored-by: sline <[email protected]> Co-authored-by: 0xsline <[email protected]> Co-authored-by: Harold Hunt <[email protected]> Co-authored-by: Shivam <[email protected]> Co-authored-by: 不做了睡大觉 <[email protected]> Co-authored-by: alexyyyander <[email protected]> Co-authored-by: alexyyyander <[email protected]> Co-authored-by: Octane0411 <[email protected]> Co-authored-by: Linux2010 <[email protected]> Co-authored-by: echoVic <[email protected]> Co-authored-by: Joseph Turian <[email protected]> Co-authored-by: turian <[email protected]> Co-authored-by: mukhtharcm <[email protected]> Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> Co-authored-by: Tony Dehnke <[email protected]> Co-authored-by: Sid-Qin <[email protected]> Co-authored-by: kevinWangSheng <[email protected]> Co-authored-by: ningding97 <[email protected]> Co-authored-by: Naylenv <[email protected]> Co-authored-by: clawbie <[email protected]> Co-authored-by: Takhoffman <[email protected]> Co-authored-by: Ayane <[email protected]> Co-authored-by: Ayane <[email protected]> Co-authored-by: StingNing <[email protected]> Co-authored-by: maweibin <[email protected]> Co-authored-by: maweibin <[email protected]> Co-authored-by: Josh Avant <[email protected]> Co-authored-by: Bin Deng <[email protected]> Co-authored-by: Byungsker <[email protected]> Co-authored-by: altaywtf <[email protected]> Co-authored-by: Altay <[email protected]> Co-authored-by: jiangnan <[email protected]> Co-authored-by: jnMetaCode <[email protected]> Co-authored-by: Jacob Riff <[email protected]> Co-authored-by: jriff <[email protected]> Co-authored-by: Naylenv <[email protected]> Co-authored-by: Bill <[email protected]> Co-authored-by: 2233admin <[email protected]> Co-authored-by: 2233admin <[email protected]> Co-authored-by: littleben <[email protected]> Co-authored-by: littleben <[email protected]> Co-authored-by: OpenClaw Agent <[email protected]> Co-authored-by: Vignesh Natarajan <[email protected]> Co-authored-by: zerone0x <[email protected]> Co-authored-by: zerone0x <[email protected]> Co-authored-by: dunamismax <[email protected]> Co-authored-by: Octane <[email protected]> Co-authored-by: Mitsuyuki Osabe <[email protected]> Co-authored-by: Tak <[email protected]> Co-authored-by: Tyler Yust <[email protected]> Co-authored-by: Hinata Kaga (samon) <[email protected]> Co-authored-by: dorukardahan <[email protected]> Co-authored-by: Tyler Yust <[email protected]> Co-authored-by: Brenner Spear <[email protected]> Co-authored-by: aerelune <[email protected]> Co-authored-by: Frank Yang <[email protected]> Co-authored-by: zhouhe-xydt <[email protected]> Co-authored-by: zhouhe-xydt <[email protected]> Co-authored-by: Vignesh <[email protected]> Co-authored-by: OpenClaw SRE Bot <[email protected]> Co-authored-by: Xinhua Gu <[email protected]> Co-authored-by: Val Alexander <[email protected]>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This pull request fixes a long-standing subagent announcement regression where parent announcements were sometimes stale, wrong, or misleading.
What was broken
assistantin places where the announcement should stay as ausermessage.What this PR fixes (5 commit breakdown)
How it works now
Testing
Notes
This also addresses regressions introduced by later PRs that unintentionally overwrote Tyler's original subagent announce fixes in
#22223 and #23166.
Thanks @tyler6204