Skip to content

feat(kb): add OpenClaw — messaging-channel gateway for LLM agents#56

Merged
jonathansantilli merged 1 commit intomainfrom
feat/kb-openclaw
Apr 22, 2026
Merged

feat(kb): add OpenClaw — messaging-channel gateway for LLM agents#56
jonathansantilli merged 1 commit intomainfrom
feat/kb-openclaw

Conversation

@jonathansantilli
Copy link
Copy Markdown
Owner

Summary

Adds OpenClaw to CodeGate's knowledge base so `inventory` picks it up like any other AI tool.

OpenClaw (openclaw.ai, github.com/openclaw/openclaw) routes LLM agents through messaging channels (WhatsApp, Telegram, Signal, Slack, iMessage, Discord, LINE, Matrix, Teams, Zalo, voice). It's a different shape of risk than a coding assistant, but squarely in CodeGate's remit:

  • Every connected channel is an attacker-controllable input to the agent — classic prompt-injection surface.
  • Plugin system installs via `npm install --omit=dev` with arbitrary package content; lifecycle scripts (postinstall, etc.) execute in-process.
  • `hooks.json5` declares shell commands the agent runs before/after tool calls.

KB entry

```
config_paths: 6 — openclaw.json, custom.json, hooks.json5,
exec-approvals.json, .env, telegram.token
skill_paths: 3 — extensions//package.json, agents//agent,
workspace/IDENTITY.md
extension_mechanisms: npm_plugin, messaging_channel, hook
```

Verification

```
$ node -e 'const kb = require("codegate-ai/dist/layer1-discovery/knowledge-base.js").loadKnowledgeBase(); console.log(kb.entries.find(e => e.tool === "openclaw"))'
tool: 'openclaw', config_paths: 6, skill_paths: 3, extension_mechanisms: 3
```

  • `npm run typecheck` ✅
  • `npm run lint` ✅
  • `npm test` ✅ 154 files / 722 tests passing

OpenClaw (openclaw.ai, github.com/openclaw/openclaw) exposes LLM agents
through WhatsApp, Telegram, Signal, Slack, iMessage, Discord, LINE,
Matrix, Teams, Zalo, and voice-call channels. Not a coding assistant,
but a materially different agentic-LLM surface worth scanning:

- Untrusted inbound messages feed LLM tool-use directly (prompt-injection
  surface on every connected channel).
- Plugin system installs via `npm install --omit=dev` with arbitrary
  package content — lifecycle scripts execute in-process.
- hooks.json5 declares shell commands run before/after agent tool calls.

Registered KB entry covers:
- config_paths (6): openclaw.json, custom.json, hooks.json5,
  exec-approvals.json, .env, telegram.token
- skill_paths (3): extensions/*/package.json (plugin manifests),
  agents/*/agent (agent configs), workspace/IDENTITY.md (persona)
- extension_mechanisms (3): npm_plugin, messaging_channel, hook
@jonathansantilli jonathansantilli merged commit 854505f into main Apr 22, 2026
16 checks passed
@jonathansantilli jonathansantilli deleted the feat/kb-openclaw branch April 22, 2026 22:00
github-actions Bot pushed a commit that referenced this pull request Apr 22, 2026
# [0.15.0](v0.14.4...v0.15.0) (2026-04-22)

### Features

* **kb:** add OpenClaw — messaging-channel gateway for LLM agents ([#56](#56)) ([854505f](854505f))
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant