fix(feishu): strip internal tool-trace banners from outbound text by ZengWen-DT · Pull Request #98705 · openclaw/openclaw

ZengWen-DT · 2026-07-01T16:07:49Z

What Problem This Solves

Fixes an issue where Feishu users receiving bot replies would see internal assistant tool-trace scaffolding lines (e.g. ⚠️ 🛠️ \search repos (agent)` failed`) leaked into outbound messages. These banners are internal runtime diagnostics that should not reach end users.

Why This Change Was Made

The Feishu outbound adapter lacked a sanitizeText hook, so core stripInternalRuntimeScaffolding was the only sanitization layer — and it does not remove tool-trace failure banners. Adding sanitizeAssistantVisibleText on the outbound sanitizeText hook matches the accepted sibling pattern already shipped for Slack, Signal, Matrix, Telegram, Google Chat, QQBot, IRC, and SMS.

User Impact

Feishu users no longer see internal tool-trace failure banners in bot replies. Normal assistant prose and markdown formatting are preserved.

Evidence

Negative control — core strip keeps banner (pre-fix behavior):

$ node --import tsx -e '
import { stripInternalRuntimeScaffolding } from "./src/infra/outbound/sanitize-text.js";
const text = "Done.\n⚠️ 🛠️ \`search repos (agent)\` failed";
console.log("Output:", JSON.stringify(stripInternalRuntimeScaffolding(text)));
console.log("Banner survives:", stripInternalRuntimeScaffolding(text).includes("⚠️"));'

Output: "Done.\n⚠️ 🛠️ `search repos (agent)` failed"
Banner survives: true

After fix — feishuPlugin.outbound.sanitizeText strips banner:

$ node --import tsx -e '
import { feishuPlugin } from "./extensions/feishu/src/channel.js";
const text = "Done.\n⚠️ 🛠️ \`search repos (agent)\` failed";
const result = feishuPlugin.outbound?.sanitizeText?.({ text, payload: { text } });
console.log("Output:", JSON.stringify(result));
console.log("Banner stripped:", !result?.includes("⚠️"));'

Output: "Done."
Banner stripped: true

Focused test:

$ node scripts/run-vitest.mjs run extensions/feishu/src/outbound-tool-trace-sanitize.test.ts

 ✓ feishu outbound sanitizeText
   ✓ strips internal tool-trace banners before outbound delivery
   ✓ preserves ordinary assistant prose while sanitizing

 Test Files  1 passed (1)
      Tests  2 passed (2)

What was not tested: No live Feishu round-trip — credentials were not available in this checkout. The bug is in the sanitization layer before network delivery, so the proof above covers the affected boundary without a live channel.

AI-assisted: Cursor (Claude). Proof run manually.

Made with Cursor

Part of openclaw#90684 Co-authored-by: Cursor <[email protected]>

clawsweeper · 2026-07-01T16:10:04Z

Codex review: needs maintainer review before merge. Reviewed July 1, 2026, 12:17 PM ET / 16:17 UTC.

Summary
The branch adds a Feishu outbound sanitizeText hook using sanitizeAssistantVisibleText and a focused regression test for stripping internal tool-trace banners.

PR surface: Source +2, Tests +20. Total +22 across 2 files.

Reproducibility: yes. source inspection gives a high-confidence reproduction path: current main Feishu lacks sanitizeText, while shared delivery only runs the assistant-visible sanitizer when a channel provides that hook. The PR body also includes terminal before/after output exercising the exact Feishu hook after the patch.

Review metrics: none identified.

Root-cause cluster
Relationship: fixed_by_candidate
Canonical: #90684
Summary: This PR is a Feishu-specific implementation candidate for the broader open sanitizer gap; sibling PRs cover other channel slices of the same missing outbound sanitizer boundary.

Members:

canonical: Feishu (and non-Discord channels) should apply sanitizeAssistantVisibleText() on outbound text #90684 - The PR body explicitly says it is part of this open cross-channel sanitizer issue, and the issue describes the Feishu missing-sanitizer path.
partial_overlap: fix(googlechat): sanitize internal tool-trace lines from outbound text (#90684) #95084 - The merged Google Chat PR applied the same tool-trace sanitizer pattern to a different channel slice of the canonical issue.
partial_overlap: fix(mattermost): strip internal tool-trace banners from outbound text #98693 - The open Mattermost PR appears to address the same tool-trace banner leak pattern for another channel.
partial_overlap: fix(outbound): sanitize assistant reasoning before delivery #84441 - The open outbound reasoning sanitizer PR overlaps the broader delivery-sanitization problem but is not a merged replacement for this Feishu-specific hook.

Proposal only: this assessment does not dispatch repair, suppress jobs, mutate sibling items, close, or merge anything.

Merge readiness
Overall: 🐚 platinum hermit
Proof: 🐚 platinum hermit
Patch quality: 🦞 diamond lobster
Result: ready for maintainer review.

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

Rank-up moves:

[P2] Wait for the remaining required CI shards to complete before merge.

Risk before merge

[P1] No live Feishu/Lark round trip was shown; the supplied proof exercises the production outbound sanitizer hook before network delivery, which is the changed boundary.

Maintainer options:

Decide the mitigation before merge
Land the narrow Feishu hook after required CI and maintainer review, then keep the broader cross-channel cleanup tracked on Feishu (and non-Discord channels) should apply sanitizeAssistantVisibleText() on outbound text #90684.
Pause or close
Do not merge this PR until maintainers decide whether the risk is worth taking.

Next step before merge

No ClawSweeper repair job is needed because the review found no actionable patch defect; remaining action is normal maintainer review and required checks.

Security
Cleared: The diff reuses an existing sanitizer hook and adds a focused test; it does not touch dependencies, workflows, secrets, permissions, package resolution, or credential handling.

Review details

Best possible solution:

Land the narrow Feishu hook after required CI and maintainer review, then keep the broader cross-channel cleanup tracked on #90684.

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

Yes, source inspection gives a high-confidence reproduction path: current main Feishu lacks sanitizeText, while shared delivery only runs the assistant-visible sanitizer when a channel provides that hook. The PR body also includes terminal before/after output exercising the exact Feishu hook after the patch.

Is this the best way to solve the issue?

Yes, this is the best narrow fix for the Feishu slice because it uses the existing plugin SDK sanitizer seam and matches accepted sibling channel implementations without adding config or a parallel sanitizer.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against a5e11c0083c5.

Label changes

Label justifications:

P1: The PR fixes internal tool/runtime trace banners leaking into a real user-facing channel, matching an urgent agent/channel workflow and privacy failure.
rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🐚 platinum hermit and patch quality is 🦞 diamond lobster.
status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (terminal): The PR body includes terminal after-fix output invoking feishuPlugin.outbound.sanitizeText on the leaked banner plus focused Vitest output; no live Feishu round trip was provided, but the proof covers the changed pre-send sanitizer boundary.
proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes terminal after-fix output invoking feishuPlugin.outbound.sanitizeText on the leaked banner plus focused Vitest output; no live Feishu round trip was provided, but the proof covers the changed pre-send sanitizer boundary.

Evidence reviewed

PR surface:

Source +2, Tests +20. Total +22 across 2 files.

View PR surface stats

Area	Files	Added	Net
Source	1	2	+2
Tests	1	20	+20
Docs	0	0	0
Config	0	0	0
Generated	0	0	0
Other	0	0	0
Total	2	22	+22

What I checked:

Current main Feishu gap: On current origin/main, feishuPlugin.outbound defines chunking and presentation behavior but has no sanitizeText hook, so Feishu cannot use the shared per-channel sanitizer stage. (extensions/feishu/src/channel.ts:1365, a5e11c0083c5)
Shared delivery sanitizer seam: The shared outbound delivery path applies handler.sanitizeText before payload normalization when a channel provides the hook, which makes the PR's Feishu hook the expected owner-boundary fix. (src/infra/outbound/deliver.ts:795, a5e11c0083c5)
Canonical sanitizer behavior: The delivery profile runs internal trace-line stripping unless explicitly disabled, and sanitizeAssistantVisibleText uses that delivery profile for outbound text. (src/shared/text/assistant-visible-text.ts:903, a5e11c0083c5)
Sibling channel pattern: Current main already uses the same sanitizer hook pattern in sibling channel plugins including Slack, Signal, Matrix, Telegram, QQBot, and Google Chat. (extensions/slack/src/channel.ts:427, a5e11c0083c5)
PR diff and regression test: The PR imports sanitizeAssistantVisibleText, adds it as Feishu's outbound sanitizeText hook, and adds a Vitest test asserting the internal banner is stripped while ordinary prose is preserved. (extensions/feishu/src/outbound-tool-trace-sanitize.test.ts:1, c9d172aac188)
Related canonical issue: The open canonical issue Feishu (and non-Discord channels) should apply sanitizeAssistantVisibleText() on outbound text #90684 tracks Feishu and other non-Discord channels missing assistant-visible sanitization; this PR is a partial Feishu implementation candidate, not a reason to close the issue before merge.

Likely related people:

steipete: Peter Steinberger appears in both Feishu channel history and shared assistant-visible sanitizer history, including Feishu runtime fixes and sanitizer leak repairs. (role: feature owner / adjacent owner; confidence: high; commits: d05e4a4bc6f2, 4c4e8a213f1a, 8ac30279b302; files: extensions/feishu/src/channel.ts, src/shared/text/assistant-visible-text.ts)
vincentkoc: Vincent Koc is blamed for the current Feishu outbound block in the shallow checkout and appears in recent Feishu and sanitizer-adjacent commit history. (role: recent area contributor; confidence: medium; commits: f588a49a3677, e7d6566b8fd1, 4a4a5968e88a; files: extensions/feishu/src/channel.ts, src/shared/text/assistant-visible-text.ts)
ZengWen-DT: Beyond authoring this PR, ZengWen-DT has a merged current-main SMS sanitizer PR using the same sibling-channel tool-trace cleanup pattern. (role: recent sibling sanitizer contributor; confidence: medium; commits: c9d172aac188; files: extensions/feishu/src/channel.ts, extensions/sms/src/send.ts)
piersonr: The outbound delivery hook surface this PR uses traces to the plugin SDK reply-payload hook work in the shared outbound delivery path. (role: outbound hook introducer; confidence: medium; commits: b474f429ee4b; files: src/infra/outbound/deliver.ts, src/channels/plugins/outbound.types.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.

* origin/main: (1287 commits) fix(android): block loopback canvas navigation (openclaw#99874) fix(hooks): suppress unhandled stdout/stderr stream errors in gmail watcher (openclaw#100519) fix(memory-core): guard qmd mcporter JSON.parse against non-JSON stdout (openclaw#98381) fix(build): fall back to tsx for build TypeScript scripts (openclaw#91262) feat(skills): suggest saving detected reusable workflows by default (openclaw#95477) (openclaw#100692) docs(changelog): remove generated release-note entries feat(telegram): offer BotFather web app flow in setup help and docs (openclaw#100540) fix(ios): unify Talk and Settings row typography on one branded detail row (openclaw#100515) feat(gateway): add persisted crash-loop breaker and fatal-config exit contract refactor(macos): lock and unify PortGuardian tunnel record persistence so concurrent app instances cannot lose orphan records (openclaw#100601) fix: stop reconnecting on protocol mismatch (openclaw#98414) fix(maint): reuse recent hosted gates after rebase (openclaw#100663) fix(ui): reopen web terminals without stale content (openclaw#100665) fix(browser): diagnose empty WSL2 Chrome replies (openclaw#100590) fix(ios): chat snaps back to bottom when scrolling to top via status-bar tap (openclaw#100502) Treat already-compacted CLI compaction as no-op (openclaw#99136) docs(changelog): remove direct main fix entry fix(feishu): strip internal tool-trace banners from outbound text (openclaw#98705) fix(message): thread --limit through to CLI formatter and surface provider pagination hints (openclaw#99089) fix(voyage): close response body stream when batch output JSONL parsing throws (openclaw#98840) ... # Conflicts: # extensions/memory-wiki/package.json

…enclaw#98705) Part of openclaw#90684 Co-authored-by: Cursor <[email protected]>

…text The Nextcloud Talk outbound adapter had no sanitizeText hook, so internal assistant tool-trace scaffolding (e.g. "⚠️ 🛠️ `exec (agent)` failed", raw <tool_call>/<function_response> XML) leaked verbatim into delivered chat messages. The same one-line sanitizeText hook was already shipped to 13 sibling channels (mattermost openclaw#98693, feishu openclaw#98705, twitch openclaw#103109, signal openclaw#97360, slack openclaw#97367, matrix openclaw#97372, irc openclaw#97214, telegram, googlechat openclaw#90684, qqbot openclaw#90132, discord, whatsapp openclaw#71830, sms). Add sanitizeText: ({ text }) => sanitizeAssistantVisibleText(text) to the outbound base adapter, matching those precedents. Co-Authored-By: Claude <[email protected]>

The Nostr outbound adapter had no sanitizeText hook, so internal assistant tool-trace scaffolding (e.g. "⚠️ 🛠️ `exec (agent)` failed", raw <tool_call>/<function_response> XML) leaked verbatim into delivered Nostr DMs. The same one-line sanitizeText hook was already shipped to 13 sibling channels (twitch openclaw#103109, mattermost openclaw#98693, feishu openclaw#98705, signal openclaw#97360, slack openclaw#97367, matrix openclaw#97372, irc openclaw#97214, telegram, googlechat openclaw#90684, qqbot openclaw#90132, discord, whatsapp openclaw#71830, sms). Add sanitizeText to the NostrOutboundAdapter type and adapter, matching those precedents. Co-Authored-By: Claude <[email protected]>

The Zalo outbound adapter had no sanitizeText hook, so internal assistant tool-trace scaffolding (e.g. "⚠️ 🛠️ `exec (agent)` failed", raw <tool_call>/<function_response> XML) leaked verbatim into delivered Zalo messages. The same one-line sanitizeText hook was already shipped to 13 sibling channels (twitch openclaw#103109, mattermost openclaw#98693, feishu openclaw#98705, signal openclaw#97360, slack openclaw#97367, matrix openclaw#97372, irc openclaw#97214, telegram, googlechat openclaw#90684, qqbot openclaw#90132, discord, whatsapp openclaw#71830, sms). Add sanitizeText: ({ text }) => sanitizeAssistantVisibleText(text) to the outbound adapter, matching those precedents. Co-Authored-By: Claude <[email protected]>

The Nostr outbound adapter had no sanitizeText hook, so internal assistant tool-trace scaffolding (e.g. "⚠️ 🛠️ `exec (agent)` failed", raw <tool_call>/<function_response> XML) leaked verbatim into delivered Nostr DMs. The same one-line sanitizeText hook was already shipped to 13 sibling channels (twitch openclaw#103109, mattermost openclaw#98693, feishu openclaw#98705, signal openclaw#97360, slack openclaw#97367, matrix openclaw#97372, irc openclaw#97214, telegram, googlechat openclaw#90684, qqbot openclaw#90132, discord, whatsapp openclaw#71830, sms). Add sanitizeText to the NostrOutboundAdapter type and adapter, matching those precedents. Co-Authored-By: Claude <[email protected]>

The Zalo outbound adapter had no sanitizeText hook, so internal assistant tool-trace scaffolding (e.g. "⚠️ 🛠️ `exec (agent)` failed", raw <tool_call>/<function_response> XML) leaked verbatim into delivered Zalo messages. The same one-line sanitizeText hook was already shipped to 13 sibling channels (twitch openclaw#103109, mattermost openclaw#98693, feishu openclaw#98705, signal openclaw#97360, slack openclaw#97367, matrix openclaw#97372, irc openclaw#97214, telegram, googlechat openclaw#90684, qqbot openclaw#90132, discord, whatsapp openclaw#71830, sms). Add sanitizeText: ({ text }) => sanitizeAssistantVisibleText(text) to the outbound adapter, matching those precedents. Co-Authored-By: Claude <[email protected]>

The Nostr outbound adapter had no sanitizeText hook, so internal assistant tool-trace scaffolding (e.g. "⚠️ 🛠️ `exec (agent)` failed", raw <tool_call>/<function_response> XML) leaked verbatim into delivered Nostr DMs. The same one-line sanitizeText hook was already shipped to 13 sibling channels (twitch openclaw#103109, mattermost openclaw#98693, feishu openclaw#98705, signal openclaw#97360, slack openclaw#97367, matrix openclaw#97372, irc openclaw#97214, telegram, googlechat openclaw#90684, qqbot openclaw#90132, discord, whatsapp openclaw#71830, sms). Add sanitizeText to the NostrOutboundAdapter type and adapter, matching those precedents. Co-Authored-By: Claude <[email protected]>

The Zalo outbound adapter had no sanitizeText hook, so internal assistant tool-trace scaffolding (e.g. "⚠️ 🛠️ `exec (agent)` failed", raw <tool_call>/<function_response> XML) leaked verbatim into delivered Zalo messages. The same one-line sanitizeText hook was already shipped to 13 sibling channels (twitch openclaw#103109, mattermost openclaw#98693, feishu openclaw#98705, signal openclaw#97360, slack openclaw#97367, matrix openclaw#97372, irc openclaw#97214, telegram, googlechat openclaw#90684, qqbot openclaw#90132, discord, whatsapp openclaw#71830, sms). Add sanitizeText: ({ text }) => sanitizeAssistantVisibleText(text) to the outbound adapter, matching those precedents. Co-Authored-By: Claude <[email protected]>

The Nostr outbound adapter had no sanitizeText hook, so internal assistant tool-trace scaffolding (e.g. "⚠️ 🛠️ `exec (agent)` failed", raw <tool_call>/<function_response> XML) leaked verbatim into delivered Nostr DMs. The same one-line sanitizeText hook was already shipped to 13 sibling channels (twitch openclaw#103109, mattermost openclaw#98693, feishu openclaw#98705, signal openclaw#97360, slack openclaw#97367, matrix openclaw#97372, irc openclaw#97214, telegram, googlechat openclaw#90684, qqbot openclaw#90132, discord, whatsapp openclaw#71830, sms). Add sanitizeText to the NostrOutboundAdapter type and adapter, matching those precedents. Co-Authored-By: Claude <[email protected]>

fix(feishu): strip internal tool-trace banners from outbound text

c9d172a

Part of openclaw#90684 Co-authored-by: Cursor <[email protected]>

openclaw-barnacle Bot added channel: feishu Channel integration: feishu size: XS labels Jul 1, 2026

openclaw-clownfish Bot merged commit 6378d5e into openclaw:main Jul 6, 2026
134 of 140 checks passed

vincentkoc mentioned this pull request Jul 6, 2026

chore(jobs): prune resolved OpenClaw inbox jobs openclaw/clownfish#221

Merged

github-actions Bot pushed a commit to Desicool/openclaw that referenced this pull request Jul 6, 2026

fix(feishu): strip internal tool-trace banners from outbound text (op…

c2b3615

…enclaw#98705) Part of openclaw#90684 Co-authored-by: Cursor <[email protected]>

This was referenced Jul 7, 2026

fix(line): strip internal tool-trace banners from outbound text #101708

Merged

fix(nextcloud-talk): strip internal tool-trace banners from outbound text #101712

Merged

Pick-cat mentioned this pull request Jul 7, 2026

fix(nostr): strip internal tool-trace banners from outbound text #101760

Open

clawsweeper Bot mentioned this pull request Jul 8, 2026

fix(imessage): apply canonical outbound sanitizer #101430

Open

masatohoshino mentioned this pull request Jul 8, 2026

fix(tlon): strip internal tool-trace banners from outbound text #102211

Closed

giodl73-repo pushed a commit to giodl73-repo/openclaw that referenced this pull request Jul 8, 2026

fix(feishu): strip internal tool-trace banners from outbound text (op…

3a44aa8

…enclaw#98705) Part of openclaw#90684 Co-authored-by: Cursor <[email protected]>

clawsweeper Bot mentioned this pull request Jul 9, 2026

fix(synology-chat): strip internal tool-trace banners from outbound text #102925

Merged

This was referenced Jul 10, 2026

fix(nextcloud-talk): strip internal tool-trace banners from outbound text #103355

Closed

fix(nostr): strip internal tool-trace banners from outbound text #103361

Open

fix(zalo): strip internal tool-trace banners from outbound text #103377

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

fix(feishu): strip internal tool-trace banners from outbound text#98705

fix(feishu): strip internal tool-trace banners from outbound text#98705
openclaw-clownfish[bot] merged 1 commit into
openclaw:mainfrom
ZengWen-DT:fix/feishu-outbound-tool-trace-sanitize

ZengWen-DT commented Jul 1, 2026

Uh oh!

clawsweeper Bot commented Jul 1, 2026 •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Uh oh!

Conversation

ZengWen-DT commented Jul 1, 2026

What Problem This Solves

Why This Change Was Made

User Impact

Evidence

Uh oh!

clawsweeper Bot commented Jul 1, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

clawsweeper Bot commented Jul 1, 2026 •

edited

Loading