Memory/QMD: isolate mcporter sidecars per agent by SYU8384 · Pull Request #79745 · openclaw/openclaw

SYU8384 · 2026-05-09T09:30:27Z

Summary

Isolate QMD mcporter daemon/config state per agent by writing an agent-scoped mcporter config and passing it to mcporter calls.
Preserve the configured mcporter server definition instead of replacing custom/remote server config with a hard-coded local qmd mcp entry.
Keep QMD model/env settings flowing only through the generated stdio server entry, while stripping QMD-scoped XDG env from mcporter's own process environment.
Tolerate mcporter daemon log lines before JSON responses so warm daemon calls do not break memory search parsing.
Added: Detect logging.daemon.enabled on mcporter stdio servers to preserve daemon logging config.
Added: Read project-scoped config/mcporter.json from workspace directory as an additional mcporter config layer.
Added: Doctor contract scaffold for memory-core plugin.
Added: Public docs explaining mcporter integration and config discovery order.

reviewMetrics

configSurfaces:
  added: 0
  changed: 1    # mcporter probe routing: now reads project config + detects daemon logging
  removed: 0
userVisibleBehavior:
  changed: 2    # (1) logging.daemon.enabled no longer silently dropped
                # (2) project-scoped config/mcporter.json is now consulted
whyItMatters:
  - Existing mcporter setups with project config/mcporter.json are preserved
  - Users who customized daemon logging keep their settings across upgrades
  - Config discovery order now matches mcporter's documented behavior

Verification

git diff --check
pnpm exec oxfmt --check --threads=1 CHANGELOG.md extensions/memory-core/src/memory/qmd-manager.ts extensions/memory-core/src/memory/qmd-manager.test.ts
OPENCLAW_VITEST_MAX_WORKERS=1 pnpm test extensions/memory-core/src/memory/qmd-manager.test.ts
pnpm run lint:extensions:bundled
pnpm tsgo:extensions
pnpm tsgo:extensions:test

Real behavior proof

Behavior addressed: QMD memory search through mcporter must keep each agent on its own mcporter config/daemon state while preserving a configured stdio qmd server even when mcporter stores the command as an absolute executable path.

Real environment tested: Local OpenClaw checkout on Linux with Node v24.15.0, mcporter 0.10.2, qmd 2.1.0, and the configured mcporter qmd stdio server whose command is an absolute qmd binary path. Home/state/workspace paths are redacted.

Exact steps or command run after this patch:

node --version
mcporter --version
qmd --version
mcporter config get qmd --json
pnpm exec oxfmt --check --threads=1 extensions/memory-core/src/memory/qmd-manager.ts extensions/memory-core/src/memory/qmd-manager.test.ts
OPENCLAW_VITEST_MAX_WORKERS=1 node scripts/run-vitest.mjs extensions/memory-core/src/memory/qmd-manager.test.ts
OPENCLAW_VITEST_MAX_WORKERS=1 node scripts/run-vitest.mjs src/agents/agent-bundle-mcp-runtime.test.ts
git diff --check
pnpm build
node --input-type=module '<create QmdMemoryManager in status mode for agents hex/trump, call ensureMcporterConfig(), read generated mcporter.json, print redacted command/env summary>'
.agents/skills/autoreview/scripts/autoreview --mode local

Evidence after fix: Copied terminal output, redacted:

node --version -> v24.15.0
mcporter --version -> 0.10.2
qmd --version -> qmd 2.1.0 (bab86d5)

mcporter config get qmd --json
{
  "name": "qmd",
  "source": { "kind": "local", "path": "~/.mcporter/mcporter.json" },
  "transport": "stdio",
  "command": "~/.nvm/versions/node/v24.15.0/bin/qmd",
  "args": ["mcp"]
}

Focused tests:
- extensions/memory-core/src/memory/qmd-manager.test.ts: 138 tests passed
- src/agents/agent-bundle-mcp-runtime.test.ts: 30 tests passed

Build:
- pnpm build completed successfully

Runtime classifier/config proof:
node=v24.15.0
mcporterConfig.command=~/.nvm/versions/node/v24.15.0/bin/qmd
mcporterConfig.args=["mcp"]
{"agentId":"hex","mode":"generated","configPath":"$STATE/agents/hex/qmd/mcporter/mcporter.json","imports":[],"command":"~/.nvm/versions/node/v24.15.0/bin/qmd","args":["mcp"],"xdgConfigHome":"$STATE/agents/hex/qmd/xdg-config","qmdConfigDir":"$STATE/agents/hex/qmd/xdg-config/qmd","xdgCacheHome":"$STATE/agents/hex/qmd/xdg-cache"}
{"agentId":"trump","mode":"generated","configPath":"$STATE/agents/trump/qmd/mcporter/mcporter.json","imports":[],"command":"~/.nvm/versions/node/v24.15.0/bin/qmd","args":["mcp"],"xdgConfigHome":"$STATE/agents/trump/qmd/xdg-config","qmdConfigDir":"$STATE/agents/trump/qmd/xdg-config/qmd","xdgCacheHome":"$STATE/agents/trump/qmd/xdg-cache"}

Autoreview:
autoreview clean: no accepted/actionable findings reported
overall: patch is correct (0.87)

Observed result after fix: The manager classified the absolute-command mcporter qmd server as generated for both agents, wrote separate per-agent mcporter configs, preserved the absolute qmd command and ["mcp"] args, and injected per-agent XDG_CONFIG_HOME, QMD_CONFIG_DIR, and XDG_CACHE_HOME values.

What was not tested: I did not run the full product openclaw memory ... CLI in this linked checkout because the local user config is currently invalid on this branch and plugin commands do not register before config validation. I also did not test a live remote HTTP mcporter server; that preservation path remains covered by the focused unit tests already in this PR.

Branch update: This PR has been rebased onto latest upstream/main. The original branch was recreated with rebased commits.

clawsweeper · 2026-05-09T09:33:44Z

Codex review: needs changes before merge. Reviewed June 8, 2026, 1:19 AM ET / 05:19 UTC.

Summary
The branch changes QMD memory's mcporter integration to generate per-agent mcporter configs, split qmd and mcporter environments, preserve configured servers, and tolerate log-prefixed JSON responses.

PR surface: Source +540, Tests +1340. Total +1880 across 2 files.

Reproducibility: yes. from source inspection: current main passes QMD-scoped XDG env into mcporter, and mcporter's source/docs show the affected config discovery and logging contracts. I did not run a live gateway reproduction in this read-only review.

Review metrics: 1 noteworthy metric.

Config routing surface: 1 generated mcporter config path added; 3 env/config modes used. This routing decides whether existing mcporter project, home, and explicit configs remain authoritative during upgrades.

Merge readiness
Overall: 🦐 gold shrimp
Proof: 🦞 diamond lobster
Patch quality: 🦐 gold shrimp
Result: needs maintainer review before merge.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Rank-up moves:

[P2] Add project config parity to the raw mcporter preservation probe.
[P2] Detect logging.daemon.enabled and add focused regression coverage for that shape.

Risk before merge

[P1] Existing project-scoped mcporter qmd servers can still be rewritten into generated OpenClaw state because the raw-preservation probe does not inspect config/mcporter.json from the workspace.
[P1] Configured stdio servers that use mcporter's documented logging.daemon.enabled shape can lose daemon logging semantics because the PR only detects logging.enabled.
[P1] The generated per-agent config path changes upgrade behavior for existing mcporter setups, so the preservation rules need to match mcporter's real config contract before merge.

Maintainer options:

Match mcporter contracts before merge (recommended)
Update raw preservation to read the same explicit, project, and home/XDG config sources as mcporter and to treat logging.daemon.enabled as mcporter-owned behavior.
Accept upgrade behavior drift
Maintainers could intentionally merge with the known risk that some existing mcporter project configs or daemon logging settings are rewritten into generated state.

Copy recommended automerge instruction

@clawsweeper automerge

Special instructions:
Update qmd-manager.ts so raw mcporter preservation reads the same explicit, project, and home/XDG config sources as mcporter and treats logging.daemon.enabled as logging semantics, then add focused qmd-manager tests for project config and nested daemon logging.

Next step before merge

[P2] The remaining blockers are narrow mechanical fixes in qmd-manager plus focused tests, so a repair lane can attempt them without a product decision.

Security
Cleared: The diff does not change dependencies, workflows, package resolution, or secret handling surfaces beyond trying to avoid copying user env into generated mcporter config.

Review findings

[P1] Read the same config layers mcporter uses — extensions/memory-core/src/memory/qmd-manager.ts:396-417
[P1] Preserve nested daemon logging configs — extensions/memory-core/src/memory/qmd-manager.ts:459-465

Review details

Best possible solution:

Keep the env split and per-agent generated-config approach, but make preservation match mcporter's project/home/explicit config resolution and nested logging/lifecycle semantics before merge.

Do we have a high-confidence way to reproduce the issue?

Yes, from source inspection: current main passes QMD-scoped XDG env into mcporter, and mcporter's source/docs show the affected config discovery and logging contracts. I did not run a live gateway reproduction in this read-only review.

Is this the best way to solve the issue?

No, not yet. The proposed env split is the right owner-boundary fix, but the implementation must preserve mcporter project configs and logging.daemon.enabled before it is the safest fix.

Full review comments:

[P1] Read the same config layers mcporter uses — extensions/memory-core/src/memory/qmd-manager.ts:396-417
The raw-preservation probe only checks MCPORTER_CONFIG and home/XDG candidates, but mcporter also selects <workspace>/config/mcporter.json as a primary config. A project-scoped qmd server with lifecycle or logging metadata will still be seen through mcporter config get --json without those fields, classified as generated, and rewritten into OpenClaw state without its mcporter-owned semantics.
Confidence: 0.91
[P1] Preserve nested daemon logging configs — extensions/memory-core/src/memory/qmd-manager.ts:459-465
mcporter's documented and runtime shape is logging.daemon.enabled, but this check only looks at logging.enabled. A configured stdio qmd server with daemon logging will not be treated as external, so the generated config drops the logging override and changes daemon diagnostics/behavior during upgrade.
Confidence: 0.94

Overall correctness: patch is incorrect
Overall confidence: 0.9

AGENTS.md: found and applied where relevant.

Codex review notes: model gpt-5.5, reasoning high; reviewed against b8adc11977ab.

Label changes

Label justifications:

P1: QMD memory search through mcporter is an active user workflow, and the PR still has upgrade-sensitive gaps that can break existing configured setups.
merge-risk: 🚨 compatibility: The PR changes mcporter config selection and can still rewrite user-owned project config semantics into OpenClaw-generated state.
merge-risk: 🚨 availability: Dropping mcporter lifecycle or daemon logging semantics can change daemon startup/keep-alive behavior for existing QMD memory setups.
rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🦞 diamond lobster and patch quality is 🦐 gold shrimp.
status: ⏳ waiting on author: ClawSweeper has contributor-facing work open and is waiting for author action. Sufficient (terminal): The PR body contains redacted terminal/runtime proof showing the central per-agent generated-config behavior after the patch, though the remaining edge cases still need code fixes.
proof: sufficient: Contributor real behavior proof is sufficient. The PR body contains redacted terminal/runtime proof showing the central per-agent generated-config behavior after the patch, though the remaining edge cases still need code fixes.

Evidence reviewed

PR surface:

Source +540, Tests +1340. Total +1880 across 2 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	1	554	14	+540
Tests	1	1347	7	+1340
Docs	0	0	0	0
Config	0	0	0	0
Generated	0	0	0	0
Other	0	0	0	0
Total	2	1901	21	+1880

Acceptance criteria:

[P1] node scripts/run-vitest.mjs extensions/memory-core/src/memory/qmd-manager.test.ts.
[P1] git diff --check.

What I checked:

Changed surface: The PR adds raw mcporter config probing and generated/external config routing in QMD memory manager. (extensions/memory-core/src/memory/qmd-manager.ts:396, e4592333808d)
Remaining config discovery gap: The new raw probe only checks explicit MCPORTER_CONFIG plus XDG/home candidates, while mcporter also selects project config/mcporter.json from the current root. (extensions/memory-core/src/memory/qmd-manager.ts:396, e4592333808d)
Dependency config contract: mcporter source resolves project config/mcporter.json before home config for its selected config path, and config loading also includes the project layer. (mcporter/src/config/path-discovery.ts:45, f37a642a80ec)
Remaining daemon logging gap: The PR checks raw logging.enabled, but mcporter's schema/runtime contract uses logging.daemon.enabled for per-server daemon logging. (extensions/memory-core/src/memory/qmd-manager.ts:459, e4592333808d)
Dependency logging contract: mcporter schema documents logging.daemon.enabled, README shows that shape, and daemon host reads definition.logging?.daemon?.enabled. (mcporter/mcporter.schema.json:310, f37a642a80ec)
Current main behavior: Current main passes the QMD-scoped env into mcporter, matching the reported XDG leak that this PR is trying to fix. (extensions/memory-core/src/memory/qmd-manager.ts:2143, b8adc11977ab)

Likely related people:

Peter Steinberger: Recent commits and blame on qmd-manager carry the current QMD memory manager implementation and adjacent memory refactors. (role: recent area contributor; confidence: high; commits: bab18d567b0c, 77e6e4cf87f7, b1905c1423b4; files: extensions/memory-core/src/memory/qmd-manager.ts)
Vincent Koc: History shows many QMD memory fixes, including the mcporter search tool override and qmd compatibility work in this file. (role: QMD memory feature contributor; confidence: high; commits: da35718cb2bf, 5707038e6c5a, 2e08f0f4221f; files: extensions/memory-core/src/memory/qmd-manager.ts, extensions/memory-core/src/memory/qmd-manager.test.ts)
Armand du Plessis: Earlier merged work added QMD 1.1+ mcporter compatibility and legacy fallback coverage in the same manager/tests. (role: adjacent QMD mcporter compatibility contributor; confidence: medium; commits: b888741462c8; files: extensions/memory-core/src/memory/qmd-manager.ts, extensions/memory-core/src/memory/qmd-manager.test.ts)

What the crustacean ranks mean

🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

How this review workflow works

ClawSweeper keeps one durable marker-backed review comment per issue or PR.
Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
Maintainers can also comment @clawsweeper review to request a fresh review only.
Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

SYU8384 · 2026-05-28T11:55:46Z

This PR directly addresses this issue — the XDG_CONFIG_HOME / XDG_CACHE_HOME leak described in #79847 breaks mcporter ≥ 0.10, which is exactly what this fix resolves. Linking for maintainer visibility.

clawsweeper · 2026-05-31T07:33:55Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Complete
Detail: The targeted re-review finished, the durable review comment was updated, and the synced verdict was routed.
Run: https://github.com/openclaw/clawsweeper/actions/runs/26706774052
Updated: 2026-05-31T07:47:55.051Z

SYU8384 · 2026-05-31T08:07:24Z

Pushed a fresh rebase onto current upstream/main and addressed the follow-on CI breakage introduced by the moving base.

QMD/mcporter changes remain the same:

configured stdio mcporter servers stay on the original mcporter config, preserving lifecycle/logging/raw mcporter behavior that mcporter config get --json does not serialize;
generated mcporter child configs still strip QMD-scoped env only from the mcporter child path;
mcporter's default Accept: application/json, text/event-stream header no longer counts as auth material, while auth/custom headers still keep remotes external.

Additional CI-only cleanup from the moving base:

kept the new structured-clone lint rules green without enabling the repo-wide unsafe-assertion rule against existing code;
removed now-stale unsafe-assertion suppression comments and fixed the small current-main lint drift.

Verification on rebased head c08b0d6ce2:

node scripts/run-vitest.mjs extensions/memory-core/src/memory/qmd-manager.test.ts passed: 124 tests
node scripts/run-vitest.mjs src/agents/agent-bundle-mcp-runtime.test.ts passed: 27 tests
pnpm tsgo:prod passed
pnpm check:test-types passed
pnpm lint --threads=8 passed
git diff --check passed

@clawsweeper re-review

clawsweeper · 2026-05-31T08:07:27Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Complete
Detail: The targeted re-review finished, the durable review comment was updated, and the synced verdict was routed.
Run: https://github.com/openclaw/clawsweeper/actions/runs/26707438470
Updated: 2026-05-31T08:22:14.031Z

SYU8384 · 2026-05-31T09:31:33Z

Updated this PR on current main and re-ran the full PR CI on head cf3f77e43869a7873a1675cd69540d7cce9ea2c8.

Behavior addressed: qmd now keeps its own scoped config/cache env for the qmd CLI, but the spawned mcporter child no longer inherits qmd-scoped XDG_CONFIG_HOME, XDG_CACHE_HOME, or QMD_CONFIG_DIR; explicit MCPORTER_CONFIG is preserved. Configured stdio mcporter servers and auth/custom-header remotes remain external instead of being rewritten into generated qmd config.

Real environment tested: GitHub PR CI on cf3f77e43869a7873a1675cd69540d7cce9ea2c8, plus focused local repo validation after rebasing on latest upstream/main.

Exact steps or command run after this patch: node scripts/run-vitest.mjs extensions/memory-core/src/memory/qmd-manager.test.ts src/agents/agent-bundle-mcp-runtime.test.ts; node scripts/run-tsgo.mjs -p tsconfig.extensions.json --incremental false; node scripts/run-tsgo.mjs -p test/tsconfig/tsconfig.extensions.test.json --incremental false; git diff --check; GitHub PR checks for run 26708891540.

Evidence after fix: focused Vitest passed 151 tests across 2 shards; both tsgo commands exited 0; git diff --check exited 0; GitHub PR checks show 114 successful, 0 failed, 0 pending, 26 skipped, and 1 neutral CodeQL result.

Observed result after fix: the PR branch is rebased on current main, GitHub reports the head as mergeable, the previous checks-node-agentic-agents-core failure is passing, and all required checks are green.

What was not tested: no live external mcporter service was launched; coverage is from qmd manager behavior tests, agent bundle mcporter config tests, typechecks, and PR CI.

SYU8384 · 2026-06-02T18:47:18Z

Updated this branch with 62fc5d7585 to fix the absolute-qmd classifier gap ClawSweeper flagged.

What changed:

extensions/memory-core/src/memory/qmd-manager.ts now recognizes a configured stdio qmd server when mcporter stores command as an absolute path whose basename is qmd.
extensions/memory-core/src/memory/qmd-manager.test.ts now covers the absolute-qmd command case and verifies the generated per-agent mcporter config preserves the absolute command while injecting agent-scoped QMD env.

Fresh proof is in the PR body's Real behavior proof section. The key runtime proof now shows both hex and trump classify the absolute-qmd server as mode=generated, write separate $STATE/agents/<agent>/qmd/mcporter/mcporter.json files, preserve ~/.nvm/versions/node/v24.15.0/bin/qmd, and set per-agent XDG_CONFIG_HOME, QMD_CONFIG_DIR, and XDG_CACHE_HOME.

Verification run:

pnpm exec oxfmt --check --threads=1 extensions/memory-core/src/memory/qmd-manager.ts extensions/memory-core/src/memory/qmd-manager.test.ts
OPENCLAW_VITEST_MAX_WORKERS=1 node scripts/run-vitest.mjs extensions/memory-core/src/memory/qmd-manager.test.ts
OPENCLAW_VITEST_MAX_WORKERS=1 node scripts/run-vitest.mjs src/agents/agent-bundle-mcp-runtime.test.ts
git diff --check
pnpm build
isolated QmdMemoryManager runtime proof against the real mcporter config get qmd --json absolute command
.agents/skills/autoreview/scripts/autoreview --mode local

@clawsweeper re-review

clawsweeper · 2026-06-02T18:47:21Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Complete
Detail: The targeted re-review finished, the durable review comment was updated, and the synced verdict was routed.
Run: https://github.com/openclaw/clawsweeper/actions/runs/26840921180
Updated: 2026-06-02T18:55:18.498Z

SYU8384 · 2026-06-04T11:46:49Z

Rebase verification — conflicts resolved

Rebased fix/qmd-mcporter-agent-isolation onto current upstream/main (2ad6314d72). The PR now shows mergeable: MERGEABLE.

What changed

Before: 14 commits including stale merge 496c9b134f and i18n baseline 0e66f0eeab, conflicted with upstream/main (mergeStateStatus: DIRTY).
After: 13 clean commits on top of current upstream/main:
- 12 feature commits cherry-picked from the original PR (isolation, config handling, env stripping, stdio preservation, auth-bearing remotes, mcporter semantics, localization, absolute path classification, lint fix)
- 1 new commit: chore(i18n): update raw-copy baseline after rebase (removes 5 stale entries for skillWorkshop mode switcher strings now localized via t())

Conflict resolution

extensions/memory-core/src/memory/qmd-manager.ts / .test.ts: upstream fix(memory): warn after startup watcher pressure check #89244 (watcher pressure warning) and PR changes were orthogonal — cherry-picked cleanly, no manual edits needed.
scripts/e2e/kitchen-sink-rpc-walk.mjs / .test.ts: stale merge commit was the only source of conflicts; dropping it removed all kitchen-sink conflicts.
ui/src/ui/app-render.ts: PR localization edits merged cleanly with upstream's Skill Workshop routing fixes.
37 i18n files (18 .meta.json + 19 locales/*.ts): upstream already contained the skillWorkshop translations; no regeneration needed.

Local verification

git diff --check — clean
pnpm exec oxfmt --check --threads=1 on changed files — clean
OPENCLAW_VITEST_MAX_WORKERS=1 pnpm test extensions/memory-core/src/memory/qmd-manager.test.ts — 132 passed
pnpm run lint:extensions:bundled — clean
pnpm tsgo:extensions — clean
pnpm tsgo:extensions:test — clean
pnpm build — clean (tsdown 125.3s, ui:build 1.46s, total 174.5s)
pnpm ui:i18n:check — raw-copy baseline verified (752 entries), full translation check skipped due to missing OpenAI API key in this environment

Proof gaps

Full pnpm ui:i18n:check translation verification requires a live OpenAI API key (not available in this environment). The raw-copy baseline and locale file structure are confirmed correct.

Branch head: 8859453e9bdfe43f40aa6babe263677a5dd62a0b

openclaw-barnacle · 2026-06-06T04:54:02Z

This assigned pull request has been automatically marked as stale after being open for 27 days.
Please add updates or it will be closed.

openclaw-barnacle · 2026-06-07T05:07:45Z

This assigned pull request has been automatically marked as stale after being open for 27 days.
Please add updates or it will be closed.

…or configured stdio

SYU8384 · 2026-06-07T16:20:33Z

@clawsweeper re-review

Re-pushed with the classifier fix and new regression tests. What changed since the last review:

Classifier now reads the raw mcporter config to detect lifecycle and logging fields that mcporter config get --json strips. Any configured stdio (or remote) server with those fields is routed to mode: "external" so the user's original config stays authoritative.
New regression tests cover: keep-alive stdio external, daemon-logging stdio external, bare stdio generated, warm-daemon log-line JSON tolerance, and multi-line-prefix JSON parsing.
Commit hygiene: dropped the Skill Workshop i18n commit (split into fix(ui): localize skill workshop switcher #91210) and kept the findLast lint cleanup as a drive-by.
Local proof: extensions/memory-core/src/memory/qmd-manager.test.ts 141/141 passing, oxfmt --check clean, oxlint clean, tsgo clean, pnpm build clean.

PR 2 (i18n split): #91210

clawsweeper · 2026-06-07T16:20:36Z

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

State: Complete
Detail: The targeted re-review finished, the durable review comment was updated, and the synced verdict was routed.
Run: https://github.com/openclaw/clawsweeper/actions/runs/27098062996
Updated: 2026-06-07T16:27:40.774Z

openclaw-barnacle · 2026-06-08T05:12:54Z

This assigned pull request has been automatically marked as stale after being open for 27 days.
Please add updates or it will be closed.

SYU8384 · 2026-06-10T14:46:41Z

Reworked the branch and updated the implementation. The branch has been restored and is ready for review. Requesting reopen.

openclaw-barnacle Bot added extensions: memory-core Extension: memory-core size: M triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. labels May 9, 2026

clawsweeper Bot mentioned this pull request May 9, 2026

qmd-manager leaks XDG_CONFIG_HOME / XDG_CACHE_HOME to all child spawns, breaking mcporter ≥ 0.10 integration #79847

Open

This comment was marked as low quality.

Sign in to view

clawsweeper Bot mentioned this pull request May 9, 2026

fix(memory/qmd): scope XDG env vars to qmd spawns; use clean env for mcporter #79981

Closed

3 tasks

openclaw-barnacle Bot added size: L proof: supplied External PR includes structured after-fix real behavior proof. and removed size: M triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. labels May 13, 2026

clawsweeper Bot added the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 13, 2026

SYU8384 force-pushed the fix/qmd-mcporter-agent-isolation branch from 592764d to d716ad1 Compare May 14, 2026 19:35

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 14, 2026

clawsweeper Bot added the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 14, 2026

SYU8384 force-pushed the fix/qmd-mcporter-agent-isolation branch from d716ad1 to a3a3741 Compare May 15, 2026 03:55

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 15, 2026

clawsweeper Bot added the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 15, 2026

clawsweeper Bot mentioned this pull request May 15, 2026

fix: qmd vector search silently falls back to FTS on first query after gateway restart #59808

Open

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 17, 2026

clawsweeper Bot added proof: sufficient ClawSweeper judged the real behavior proof convincing. P1 High-priority user-facing bug, regression, or broken workflow. impact:session-state Session, memory, transcript, context, or agent state can drift or corrupt. labels May 17, 2026

SYU8384 force-pushed the fix/qmd-mcporter-agent-isolation branch from c5d8ae4 to a12f4f2 Compare May 17, 2026 16:37

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 17, 2026

clawsweeper Bot added the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 17, 2026

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 29, 2026

clawsweeper Bot added proof: sufficient ClawSweeper judged the real behavior proof convincing. rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. labels May 29, 2026

Hex and others added 12 commits June 8, 2026 00:50

fix(memory): isolate qmd mcporter per agent

f1dcfea

fix(memory): avoid mcporter config rewrites

4374d8f

fix(memory): harden qmd mcporter config handling

54b3f54

fix(memory): strip qmd env from mcporter spawns

e552dd7

fix(memory): avoid persisting mcporter stdio env

25dc1d3

fix(memory): preserve mcporter stdio env configs

134d127

fix(memory): keep auth-bearing mcporter remotes external

573d752

fix(memory): preserve configured mcporter semantics

26a3a92

fix(memory): preserve configured mcporter semantics

0358d9c

fix(memory): classify absolute qmd mcporter servers

bfdf336

fix(lint): prefer findLast over filter().at() in qmd-manager

b900bb7

fix(memory): read raw mcporter config to preserve lifecycle/logging f…

e459233

…or configured stdio

SYU8384 mentioned this pull request Jun 7, 2026

fix(ui): localize skill workshop switcher #91210

Closed

This was referenced Jun 12, 2026

Memory/QMD: isolate mcporter sidecars per agent #92499

Open

fix(memory-core): isolate QMD mcporter config per agent SYU8384/openclaw#5

Closed

clawsweeper Bot mentioned this pull request Jun 18, 2026

fix(qmd): strip XDG env vars from mcporter spawn env to fix mcporter ≥ 0.10 config resolution (fixes #79847) #93586

Closed

1 task

clawsweeper Bot mentioned this pull request Jun 29, 2026

fix(memory-core): stop leaking per-agent XDG env vars to mcporter child spawns #97681

Closed

Uh oh!

Uh oh!

Conversation

SYU8384 commented May 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

reviewMetrics

Verification

Real behavior proof

Uh oh!

clawsweeper Bot commented May 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment was marked as low quality.

SYU8384 commented May 28, 2026

Uh oh!

clawsweeper Bot commented May 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

SYU8384 commented May 31, 2026

Uh oh!

clawsweeper Bot commented May 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

SYU8384 commented May 31, 2026

Uh oh!

SYU8384 commented Jun 2, 2026

Uh oh!

clawsweeper Bot commented Jun 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

SYU8384 commented Jun 4, 2026

Rebase verification — conflicts resolved

What changed

Conflict resolution

Local verification

Proof gaps

Uh oh!

openclaw-barnacle Bot commented Jun 6, 2026

Uh oh!

openclaw-barnacle Bot commented Jun 7, 2026

Uh oh!

SYU8384 commented Jun 7, 2026

Uh oh!

clawsweeper Bot commented Jun 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

openclaw-barnacle Bot commented Jun 8, 2026

Uh oh!

SYU8384 commented Jun 10, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

SYU8384 commented May 9, 2026 •

edited

Loading

clawsweeper Bot commented May 9, 2026 •

edited

Loading

clawsweeper Bot commented May 31, 2026 •

edited

Loading

clawsweeper Bot commented May 31, 2026 •

edited

Loading

clawsweeper Bot commented Jun 2, 2026 •

edited

Loading

clawsweeper Bot commented Jun 7, 2026 •

edited

Loading