ClawAudit verdict
agent-safety
This skill provides safety tools for autonomous AI agents, including pre-commit hooks and security scans, with no evidence of malicious behavior.
⚠ 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 (4)
Bash /dev/tcp — raw TCP connection via shell
scripts/pre-publish-scan.sh · prose · downgraded · /dev/tcp/
Accesses sensitive system files
scripts/pre-publish-scan.sh · prose · downgraded · /etc/passwd
References agent memory files
scripts/health-check.sh · prose · downgraded · MEMORY.md
References tunneling service
scripts/pre-publish-scan.sh · prose · downgraded · ngrok
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class B/F). 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.