@clawling/clawchat-plugin-openclaw@2026.10.13-1
OpenClaw ClawChat channel plugin
AI Security Review
scanned 2d ago · by lpm-firewall-aiThe package is not clearly malicious, but its default runtime behavior can create and publicly expose a demo app using the user's account without a separate request. This is an unrequested agent capability, so downgrade to a warning.
Decision evidence
public snapshot- The `livewareSample` setting defaults to `true` when unset.
- Runtime startup launches the Liveware Sample supervisor without waiting for a user command.
- With a ClawChat token, the supervisor downloads the sample and skips only when the account already has Liveware apps.
- The default flow creates a “Liveware Sample” app, binds a public tunnel, and registers the app with ClawChat.
- The plugin can download a Liveware CLI at startup and mark it executable.
- The manifest has no install, preinstall, or postinstall hook; `prepublishOnly` is a publishing check.
- The sample bootstrap skips when the account already has Liveware apps.
- The sample download checks each file against a SHA-256 value from its manifest.
Source & flagged code
4 flaggedPackage source references dynamic require/import behavior.
dist/src/plugin-report.jsView on unpkg · L17A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.
dist/src/dm-recall-default.js#virtual:normalized:round1View on unpkgFindings
Affected versions and remediation
This report applies to @clawling/clawchat-plugin-openclaw@2026.10.13-1.
See all versions of @clawling/clawchat-plugin-openclaw for other recorded verdicts.
- Review the evidence and your use of @clawling/clawchat-plugin-openclaw@2026.10.13-1 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.