ClawAudit verdict
x-bookmark-triage
Triages X/Twitter bookmarks into knowledge cards posted to user-configured Discord using Claude Haiku; uses X OAuth 2.0 and Anthropic API key for stated purpose, no unexpected exfiltration.
⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.
Automated static analysis — not a human review. ClawAudit flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.
Findings (7)
References child_process — can spawn system processes
references/oauth-setup.md · code · child_process
Instruction-prose smuggling shape detected: collects a sensitive target ("API key") and emits it outward ("Send"). Phrased as prose with no trigger tokens — a semantic prompt-injection / data-exfil pattern the syntactic scanners can't see. Final tier capped at Caution; review the instructions before installing.
SKILL.md · | Requirement | Notes | |-------------|-------| | X OAuth 2.0 app | [developer.x.com](https://developer.x.com) — Free tier works | | Scopes | `bookmark.read` (r
Accesses sensitive environment variables
references/oauth-setup.md · code · process.env.X_OAUTH
Instructs covert action — may act without user awareness
scripts/x-oauth2-authorize.js · prose · downgraded · silently
Popular HTTP library — network access
scripts/backlog-sweep.js · prose · downgraded · Got
Node http/https module — low-level network access
scripts/triage-url.js · prose · downgraded · require('https')
References webhook/callback URL
scripts/x-oauth2-authorize.js · prose · downgraded · CALLBACK_URL
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class B/D). Final tier capped at Caution — cannot be lifted by any downgrade, example-payload opt-in, or allowlist.
Permissions & capabilities
No declared permissions — minimal attack surface.
network_in Is this flag fair?
Thanks — recorded.