fix: tool summaries preserve emoji truncation boundaries by ZengWen-DT · Pull Request #98644 · openclaw/openclaw

ZengWen-DT · 2026-07-01T13:39:19Z

What Problem This Solves

Fixes an issue where users viewing tool catalogs or /tools verbose output could see a replacement character when a tool description was truncated at the middle of an emoji or other UTF-16 surrogate pair.

This affects compact tool summaries in the effective tool inventory and Gateway tools catalog, plus verbose tool descriptions shown in /tools verbose.

Why This Change Was Made

The tool description summary helper now uses the existing UTF-16-safe truncation helper before appending ellipses. This keeps the existing whitespace and word-boundary behavior while avoiding dangling surrogate halves in user-visible tool text.

AI-assisted.

User Impact

Tool descriptions that contain emoji or other astral Unicode characters now truncate cleanly instead of rendering broken replacement characters in tool inventory surfaces.

Evidence

Before/premise:

The current raw slice(0, maxLen - 3) behavior can split a surrogate pair. A negative-control run temporarily restored raw slice in src/agents/tool-description-summary.ts and ran the new regression test:

FAIL src/agents/tool-description-summary.test.ts > keeps compact summaries UTF-16 safe at truncation boundaries
Expected: "abcd..."
Received: "abcd�..."

FAIL src/agents/tool-description-summary.test.ts > keeps verbose descriptions UTF-16 safe at truncation boundaries
Expected: "abcd..."
Received: "abcd�..."

After fix, I also ran a local OpenClaw tool-catalog proof with temporary plugin-style tool entries whose descriptions place 😀 exactly on the default truncation boundaries used by Gateway tools.catalog summaries and /tools verbose descriptions:

node --import tsx --input-type=module <<'EOF'
import { buildToolsMessage } from './src/auto-reply/status.ts';
import { summarizeToolDescriptionText } from './src/agents/tool-description-summary.ts';

const hasDanglingSurrogate = (value) => /[\uD800-\uDBFF](?![\uDC00-\uDFFF])|(?<![\uD800-\uDBFF])[\uDC00-\uDFFF]/u.test(value);
const catalogRawDescription = `${'a'.repeat(116)}😀${'b'.repeat(20)}`;
const verboseRawDescription = `${'a'.repeat(316)}😀${'b'.repeat(20)}`;
const catalogDescription = summarizeToolDescriptionText({ rawDescription: catalogRawDescription });
const toolsVerbose = buildToolsMessage({
  profile: 'full',
  groups: [
    {
      id: 'proof',
      label: 'Proof tools',
      source: 'plugin',
      tools: [
        {
          id: 'emoji_catalog_boundary_tool',
          label: 'Emoji Catalog Boundary Tool',
          description: catalogDescription,
          rawDescription: catalogRawDescription,
          source: 'plugin',
          pluginId: 'proof-tools',
        },
        {
          id: 'emoji_verbose_boundary_tool',
          label: 'Emoji Verbose Boundary Tool',
          description: 'Verbose proof tool',
          rawDescription: verboseRawDescription,
          source: 'plugin',
          pluginId: 'proof-tools',
        },
      ],
    },
  ],
}, { verbose: true });
const verboseLine = toolsVerbose.split('\n').find((line) => line.includes('Emoji Verbose Boundary Tool')) ?? '';

console.log('Gateway tools.catalog description projection:');
console.log(JSON.stringify({ id: 'emoji_catalog_boundary_tool', description: catalogDescription }, null, 2));
console.log(`catalog contains replacement char: ${catalogDescription.includes('�')}`);
console.log(`catalog contains dangling surrogate: ${hasDanglingSurrogate(catalogDescription)}`);
console.log('\n/tools verbose output excerpt:');
console.log(verboseLine);
console.log(`verbose contains replacement char: ${verboseLine.includes('�')}`);
console.log(`verbose contains dangling surrogate: ${hasDanglingSurrogate(verboseLine)}`);
EOF

Gateway tools.catalog description projection:
{
  "id": "emoji_catalog_boundary_tool",
  "description": "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa..."
}
catalog contains replacement char: false
catalog contains dangling surrogate: false

/tools verbose output excerpt:
  Emoji Verbose Boundary Tool - aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa...
verbose contains replacement char: false
verbose contains dangling surrogate: false

Focused tests and checks:

node scripts/run-vitest.mjs src/agents/tool-description-summary.test.ts src/auto-reply/status.tools.test.ts src/gateway/server-methods/tools-catalog.test.ts

Test Files  1 passed (1)
Tests  2 passed (2)
Test Files  2 passed (2)
Tests  14 passed (14)
Test Files  1 passed (1)
Tests  6 passed (6)

Additional checks:

pnpm format:check -- src/agents/tool-description-summary.ts src/agents/tool-description-summary.test.ts
All matched files use the correct format.

OPENCLAW_OXLINT_SKIP_PREPARE=1 node scripts/run-oxlint.mjs src/agents/tool-description-summary.ts src/agents/tool-description-summary.test.ts
passed

.agents/skills/autoreview/scripts/autoreview --mode local --base upstream/main
autoreview clean: no accepted/actionable findings reported

The standard oxlint wrapper without OPENCLAW_OXLINT_SKIP_PREPARE=1 was also attempted, but current upstream/main failed while preparing unrelated extension package-boundary artifacts before linting these files (plugin-sdk boundary root shims / rolldown TypeError: Cannot convert undefined or null to object). The scoped wrapper lint above still ran against the changed files.

clawsweeper · 2026-07-01T13:55:25Z

Codex review: needs maintainer review before merge. Reviewed July 1, 2026, 10:59 AM ET / 14:59 UTC.

Summary
The PR changes tool description truncation to use the existing UTF-16-safe helper and adds regression tests for compact and verbose emoji-boundary truncation.

PR surface: Source +1, Tests +32. Total +33 across 2 files.

Reproducibility: yes. from source: current main uses raw UTF-16 slice immediately before appending ellipses in both compact and verbose tool-summary paths, so an emoji can be cut between surrogate halves. I did not run a live current-main repro in this read-only review.

Review metrics: none identified.

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:

none.

Next step before merge

No ClawSweeper repair lane is needed because the submitted PR already contains the focused code and regression-test change.

Security
Cleared: No dependency, lockfile, workflow, permission, secret-handling, or third-party code-execution surface is changed.

Review details

Best possible solution:

Land the focused shared-helper fix after normal maintainer review on the exact head, keeping all existing truncation, whitespace, and word-boundary semantics except the surrogate-safe boundary.

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

Yes from source: current main uses raw UTF-16 slice immediately before appending ellipses in both compact and verbose tool-summary paths, so an emoji can be cut between surrogate halves. I did not run a live current-main repro in this read-only review.

Is this the best way to solve the issue?

Yes. Reusing the existing shared truncateUtf16Safe helper at the two raw-slice sites is the narrowest maintainable fix and avoids duplicating Unicode boundary logic.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against f35fbc898ccf.

Label changes

Label changes:

add proof: sufficient: Contributor real behavior proof is sufficient. The PR body now includes after-fix terminal output exercising the affected formatter path with emoji-boundary descriptions and showing no replacement character or dangling surrogate.
add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🐚 platinum hermit and patch quality is 🦞 diamond lobster.
add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (terminal): The PR body now includes after-fix terminal output exercising the affected formatter path with emoji-boundary descriptions and showing no replacement character or dangling surrogate.
remove rating: 🦪 silver shellfish: Current PR rating is rating: 🐚 platinum hermit, so this older rating label is no longer current.
remove status: 📣 needs proof: Current PR status label is status: 👀 ready for maintainer look.

Label justifications:

P3: This is a focused user-visible formatting bug in tool summary text with limited blast radius and no broken core workflow.
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 now includes after-fix terminal output exercising the affected formatter path with emoji-boundary descriptions and showing no replacement character or dangling surrogate.
proof: sufficient: Contributor real behavior proof is sufficient. The PR body now includes after-fix terminal output exercising the affected formatter path with emoji-boundary descriptions and showing no replacement character or dangling surrogate.

Evidence reviewed

PR surface:

Source +1, Tests +32. Total +33 across 2 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	1	3	2	+1
Tests	1	32	0	+32
Docs	0	0	0	0
Config	0	0	0	0
Generated	0	0	0	0
Other	0	0	0	0
Total	2	35	2	+33

What I checked:

Current compact truncation can split surrogates: Current main still truncates compact tool summaries with raw value.slice(0, maxLen - 3) immediately before appending ellipses. (src/agents/tool-description-summary.ts:17, f35fbc898ccf)
Current verbose truncation can split surrogates: Current main still truncates /tools verbose descriptions with raw normalized.slice(0, maxLen - 3). (src/agents/tool-description-summary.ts:139, f35fbc898ccf)
Patch uses existing safe helper: The PR head imports truncateUtf16Safe and uses it at both raw truncation sites while leaving the whitespace and word-boundary decisions unchanged. (src/agents/tool-description-summary.ts:18, 85572835ae46)
Safe helper contract: truncateUtf16Safe delegates to sliceUtf16Safe, whose boundary logic backs off when an end index would leave a dangling surrogate half. (src/shared/utf16-slice.ts:46, f35fbc898ccf)
Affected caller surfaces: The shared helper feeds effective tool inventory, Gateway tools.catalog, and /tools verbose, so fixing the shared helper is not a one-sided caller patch. (src/gateway/server-methods/tools-catalog.ts:137, f35fbc898ccf)
Real behavior proof updated: The PR body now includes copied terminal output that runs the product formatter path for /tools verbose and the catalog description projection with emoji placed at truncation boundaries, reporting no replacement character and no dangling surrogate. (85572835ae46)

Likely related people:

Takhoffman: Merged PR feat: add /tools runtime availability view #54088 added the /tools runtime availability view and introduced src/agents/tool-description-summary.ts with the raw truncation behavior. (role: introduced behavior; confidence: high; commits: 9c7823350bf6; files: src/agents/tool-description-summary.ts, src/auto-reply/status.ts, src/gateway/server-methods/tools-catalog.ts)
Bartok9: Merged PR fix(agents): truncate tool-display detail on code-point boundaries #96958 added the dependency-free UTF-16-safe slice helper for the same surrogate-boundary truncation class in tool display code. (role: adjacent invariant contributor; confidence: medium; commits: e09b9dfc1ba9; files: src/shared/utf16-slice.ts, src/agents/tool-display-common.ts)
GautamKumarOffical: Merged PR fix: correct tautological uppercase check in tool description summarizer #93753 recently fixed another bug in the same tool description summarizer. (role: recent area contributor; confidence: medium; commits: 01d3812ea24c; files: src/agents/tool-description-summary.ts)
vincentkoc: GitHub PR metadata shows this person merged recent fixes in this exact helper and the adjacent UTF-16 helper path. (role: recent merger and adjacent reviewer signal; confidence: medium; commits: 01d3812ea24c, e09b9dfc1ba9; files: src/agents/tool-description-summary.ts, src/shared/utf16-slice.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.

ZengWen-DT · 2026-07-01T14:26:24Z

@clawsweeper re-review

clawsweeper · 2026-07-01T14:26:27Z

🦞🧹
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.

ZengWen-DT · 2026-07-01T14:29:53Z

@clawsweeper re-run

…cript truncation UTF-16 safe normalizeCliContextValue (cli-output.ts) and formatSkillExperienceReviewTranscript (experience-review-prompt.ts) both used raw .slice(0, N) / .slice(-N) to truncate user-facing strings. When a truncation boundary falls inside a surrogate pair (emoji), the resulting string contains a lone surrogate — encodeURIComponent throws URIError, JSON serialisers may reject it, and downstream display surfaces can render garbage. Replace .slice(0, 200) with truncateUtf16Safe in normalizeCliContextValue. Replace .slice(0, 6_000) with truncateUtf16Safe and full.slice(-tailBudget) with sliceUtf16Safe(full, tailStart) in formatSkillExperienceReviewTranscript. Follows the pattern established by openclaw#98644, openclaw#102470, openclaw#101728 and the broader UTF-16 safety sweep already applied across the codebase.

…cript truncation UTF-16 safe (#106370) * fix(agents,skills): keep session-identity and experience-review transcript truncation UTF-16 safe normalizeCliContextValue (cli-output.ts) and formatSkillExperienceReviewTranscript (experience-review-prompt.ts) both used raw .slice(0, N) / .slice(-N) to truncate user-facing strings. When a truncation boundary falls inside a surrogate pair (emoji), the resulting string contains a lone surrogate — encodeURIComponent throws URIError, JSON serialisers may reject it, and downstream display surfaces can render garbage. Replace .slice(0, 200) with truncateUtf16Safe in normalizeCliContextValue. Replace .slice(0, 6_000) with truncateUtf16Safe and full.slice(-tailBudget) with sliceUtf16Safe(full, tailStart) in formatSkillExperienceReviewTranscript. Follows the pattern established by #98644, #102470, #101728 and the broader UTF-16 safety sweep already applied across the codebase. * chore: re-trigger CI after PR body update * test(agents,skills): construct transcripts exceeding 60k chars, verify old slices dangle Both experience-review fixtures now exceed EXPERIENCE_REVIEW_MAX_TRANSCRIPT_CHARS so the truncation branch actually fires. Pre-condition assertions prove the old raw .slice(0,6000) / .slice(tailStart) produce isolated surrogates; the production functions (formatSkillExperienceReviewTranscript, formatCliOutputError) do not. cli-output test also adds pre-condition check showing normalizeCliContextValue with raw .slice(0,200) would split a surrogate pair. * test: tighten UTF-16 truncation coverage Co-authored-by: 毛宇豪0668001457 <[email protected]> --------- Co-authored-by: Peter Steinberger <[email protected]>

…cript truncation UTF-16 safe (openclaw#106370) * fix(agents,skills): keep session-identity and experience-review transcript truncation UTF-16 safe normalizeCliContextValue (cli-output.ts) and formatSkillExperienceReviewTranscript (experience-review-prompt.ts) both used raw .slice(0, N) / .slice(-N) to truncate user-facing strings. When a truncation boundary falls inside a surrogate pair (emoji), the resulting string contains a lone surrogate — encodeURIComponent throws URIError, JSON serialisers may reject it, and downstream display surfaces can render garbage. Replace .slice(0, 200) with truncateUtf16Safe in normalizeCliContextValue. Replace .slice(0, 6_000) with truncateUtf16Safe and full.slice(-tailBudget) with sliceUtf16Safe(full, tailStart) in formatSkillExperienceReviewTranscript. Follows the pattern established by openclaw#98644, openclaw#102470, openclaw#101728 and the broader UTF-16 safety sweep already applied across the codebase. * chore: re-trigger CI after PR body update * test(agents,skills): construct transcripts exceeding 60k chars, verify old slices dangle Both experience-review fixtures now exceed EXPERIENCE_REVIEW_MAX_TRANSCRIPT_CHARS so the truncation branch actually fires. Pre-condition assertions prove the old raw .slice(0,6000) / .slice(tailStart) produce isolated surrogates; the production functions (formatSkillExperienceReviewTranscript, formatCliOutputError) do not. cli-output test also adds pre-condition check showing normalizeCliContextValue with raw .slice(0,200) would split a surrogate pair. * test: tighten UTF-16 truncation coverage Co-authored-by: 毛宇豪0668001457 <[email protected]> --------- Co-authored-by: Peter Steinberger <[email protected]>

fix(agents): preserve UTF-16 tool summary truncation

8557283

openclaw-barnacle Bot added agents Agent runtime and tooling size: XS labels Jul 1, 2026

vincentkoc merged commit d9d6108 into openclaw:main Jul 1, 2026
170 of 181 checks passed

github-actions Bot mentioned this pull request Jul 1, 2026

📡 Upstream Digest — 2026-07-01 16:19 UTC curtismercier/openclaw-mods#1182

Open

This was referenced Jul 1, 2026

fix(active-memory): keep recall summaries UTF-16 safe at truncation boundary #98711

Closed

fix(agents): keep bounded exec output UTF-16 safe #98721

Merged

github-actions Bot pushed a commit to Desicool/openclaw that referenced this pull request Jul 2, 2026

fix(agents): preserve UTF-16 tool summary truncation (openclaw#98644)

4708c74

chenyangjun-xy pushed a commit to chenyangjun-xy/openclaw that referenced this pull request Jul 3, 2026

fix(agents): preserve UTF-16 tool summary truncation (openclaw#98644)

2d4dc68

vincentkoc mentioned this pull request Jul 6, 2026

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

Merged

sheyanmin pushed a commit to sheyanmin/openclaw that referenced this pull request Jul 8, 2026

fix(agents): preserve UTF-16 tool summary truncation (openclaw#98644)

dfbcec1

Monkey-wusky mentioned this pull request Jul 13, 2026

fix(agents,skills): keep session-identity and experience-review transcript truncation UTF-16 safe #106370

Merged

Monkey-wusky mentioned this pull request Jul 15, 2026

fix(acpx): keep Pi session search term truncation UTF-16 safe #107896

Closed

clawsweeper Bot mentioned this pull request Jul 17, 2026

fix(auto-reply): preserve emoji in bash session snippets #109468

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

fix: tool summaries preserve emoji truncation boundaries#98644

fix: tool summaries preserve emoji truncation boundaries#98644
vincentkoc merged 1 commit into
openclaw:mainfrom
ZengWen-DT:fix/tool-summary-utf16-pr

ZengWen-DT commented Jul 1, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented Jul 1, 2026 •

edited

Loading

Uh oh!

ZengWen-DT commented Jul 1, 2026

Uh oh!

clawsweeper Bot commented Jul 1, 2026 •

edited

Loading

Uh oh!

ZengWen-DT commented Jul 1, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Uh oh!

Conversation

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

Uh oh!

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!

ZengWen-DT commented Jul 1, 2026

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!

ZengWen-DT commented Jul 1, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

ZengWen-DT commented Jul 1, 2026 •

edited

Loading

clawsweeper Bot commented Jul 1, 2026 •

edited

Loading

clawsweeper Bot commented Jul 1, 2026 •

edited

Loading