ClawAudit verdict
auto-research
Conducts web research via search APIs and stores results locally in an Obsidian vault and Qdrant vector store; all operations are legitimate research automation with no suspicious data routing.
⚠ 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 (3)
Possible hardcoded credential
research.sh · prose · downgraded · API_KEY="${BRAVE_API_KEY:-BSAfZrm_28TmR5FM9FhMCrTA1A3zS2n}
HTTP request to bare IP address — common in malicious payloads
SKILL.md · prose · downgraded · http://10.0.0.120
Python os.getenv — reads environment variable
vectorize.sh · prose · downgraded · os.getenv(
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class B). 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.
Is this flag fair?
Thanks — recorded.