fix(agent): emit assistant events for webchat message content by Nachx639 · Pull Request #27 · openclaw/openclaw

Nachx639 · 2025-12-13T21:22:12Z

Summary

Fix webchat receiving empty messages by passing onAgentEvent callback to runCommandReply in agentCommand

Problem

The webchat was showing "chat timed out" or empty responses because:

agentCommand calls runCommandReply but didn't pass the onAgentEvent callback
Without this callback, assistant text events (stream: "assistant") were never emitted
The gateway accumulates assistant text in chatRunBuffers and includes it in the chat event with state: "final"
Since no assistant events were emitted, the buffer was empty and message was undefined

Solution

Pass onAgentEvent to runCommandReply so that assistant stream events are properly emitted:

onAgentEvent: (evt) => {
  emitAgentEvent({
    runId: sessionId,
    stream: evt.stream,
    data: evt.data,
  });
},

Test plan

Send message via webchat → response now shows correctly
WhatsApp continues to work as before

🤖 Generated with Claude Code

The webchat was receiving empty messages because agentCommand didn't pass the onAgentEvent callback to runCommandReply. This caused assistant text events to never be emitted, resulting in the gateway broadcasting chat events with no message content. This fix passes onAgentEvent to runCommandReply so that assistant stream events are properly emitted and the webchat receives the actual response text. 🤖 Generated with [Claude Code](https://claude.ai/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>

The webchat sends an idempotencyKey with chat.send requests and expects events to arrive with that same runId. Previously, the gateway was using the internal sessionId as runId, causing a mismatch that prevented the webchat from correlating responses with pending requests. Changes: - chatRunSessions now stores {sessionKey, clientRunId} instead of just sessionKey - Chat events use clientRunId so webchat can match them to pending promises - Response payload includes the original clientRunId from idempotencyKey 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

steipete · 2025-12-14T01:07:35Z

Appreciate the fix! I went all choa-monkey on this while working on 2.0, so this was fixed without reviewing the PR in time.
Once the project stabilizes a bit I'll defo be better - thanks for the effort and testing!

feat(ui): add category filtering to command palette

…, Attack openclaw#27, Threat #7, Hardening #12)

Navbar, admin panel, and CRUD for users/orgs/members/harbors

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

- 支持根据 bindings 配置将不同用户/群聊分发到不同 Agent - 优化权限缺失时的错误提示，降低配置排障门槛 - 保留原有 accountId 默认值和流式卡片标题逻辑 Co-authored-by: wsbjj <94124135+wsbjj@users.noreply.github.com>

Memory files (MEMORY.md, memory/*.md) are the most trusted context source for agents, yet had no injection detection. External content (webhooks, emails) already uses detectSuspiciousPatterns() — this extends the same protection to memory files. Change: in manager-sync-ops.ts, call detectSuspiciousPatterns() on file content before indexFile() for memory-source files. Logs a warn with path + pattern count if suspicious content is found. Non-blocking. Fixes improvement item openclaw#27 from openclaw-improvement-ideas.md.

- openclaw#6 Reply classifier hallucination pattern: regex plus specifique (exige prenom+nom) - openclaw#7 Fuzzy match domaine: ajoute verification local part + nom similaire - openclaw#9 callOpenAI duplique: delegue au shared-nlp (tracking budget) - openclaw#22 Collision cron 14h: mini-cycles decales 10h/12h/15h/17h - openclaw#24 Bounce rate dilue: remplace derniere entree success par bounce - openclaw#26 API chat+HITL sans auth: verification x-api-token/authorization - openclaw#27 Passwords dashboard: generes aleatoirement (24 chars) - Bonus: commande "statut systeme" protegee admin-only 28/28 bugs de l'audit corriges. 0 restant. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Release: OG image and social meta tags

- Appendix: Integration Test Evidence section with Swim 7 scorecard - Evidence locations table (silas-likes-to-watch PR openclaw#27, gateway logs, canary tag) - Methodology note: 4-agent test topology, blind enrichment findings - Key finding: silent enrichment indistinguishable from training knowledge - Scrub verified: zero prince names/hostnames/IPs in main RFC body - Changelog section (marked for removal) contains internal refs as expected

Structured scorecard: 10 pass, 2 deferred, 0 regressions. Key evidence lines from raw gateway log. Enrichment trust boundary findings. Evidence archive links to karmaterminal/silas-likes-to-watch PR openclaw#27.

- Gateway log excerpts for each Swim 7 test (tolerance, width, chain, enrichment) - Each excerpt shows the guard/lifecycle event firing with timestamps - Matches evidence in silas-likes-to-watch PR openclaw#27 raw logs

…ron Filter - README.md: 24→26 active patches, added entries for both - verify-patches.sh: added checks (26/26 passing)

contributor and others added 2 commits December 13, 2025 22:53

Nachx639 force-pushed the fix/webchat-agent-events branch from 520bd0a to ead932f Compare December 13, 2025 21:53

steipete closed this Dec 14, 2025

dgarson referenced this pull request in dgarson/clawdbot Feb 2, 2026

Merge pull request #27 from dgarson/autodev-command-palette-categories

fad4fce

feat(ui): add category filtering to command palette

alexprime1889-prog pushed a commit to alexprime1889-prog/moltbot that referenced this pull request Feb 8, 2026

Fix openclaw#27: Add cache bust timestamp to force fresh Docker build

d745b34

centminmod added a commit to centminmod/clawdbot that referenced this pull request Feb 9, 2026

docs: document persistent .md file security risks across 8 docs (Gap #4…

4ee3ed0

…, Attack openclaw#27, Threat #7, Hardening #12)

frodo-harborbot pushed a commit to harborworks/openclaw that referenced this pull request Feb 16, 2026

Merge pull request openclaw#27 from harborworks/feature/navbar

f0381ba

Navbar, admin panel, and CRUD for users/orgs/members/harbors

songliu0403-rgb pushed a commit to songliu0403-rgb/openclaw that referenced this pull request Feb 26, 2026

chore: remove openclaw submodule (openclaw#27)

badc424

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

ddupg pushed a commit to ddupg/openclaw that referenced this pull request Feb 27, 2026

Merge PR openclaw#27: 飞书多 Agent 路由支持

a1f430f

xXMrNidaXx mentioned this pull request Feb 28, 2026

[Bug]: HTTP 400: tools.27.custom.input_schema: JSON schema is invalid. It must match JSON Schema draft #29416

Open

ashclaw5381089 mentioned this pull request Mar 5, 2026

feat(discord): support permission_overwrites for channel-create/channel-edit #26209

Open

vincentkoc mentioned this pull request Mar 6, 2026

Plugins: avoid false integrity drift prompts on unpinned updates #37179

Merged

18 tasks

elliot-ylambda pushed a commit to elliot-ylambda/magister-openclaw that referenced this pull request Mar 7, 2026

Merge pull request openclaw#27 from elliot-ylambda/development

86346e8

Release: OG image and social meta tags

ronan-dandelion-cult mentioned this pull request Mar 7, 2026

docs(RFC): integration test evidence appendix + scrub verification karmaterminal/openclaw#14

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix(agent): emit assistant events for webchat message content#27

fix(agent): emit assistant events for webchat message content#27
Nachx639 wants to merge 2 commits intoopenclaw:mainfrom
Nachx639:fix/webchat-agent-events

Nachx639 commented Dec 13, 2025

Uh oh!

steipete commented Dec 14, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

Nachx639 commented Dec 13, 2025

Summary

Problem

Solution

Test plan

Uh oh!

steipete commented Dec 14, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants