ClawAudit verdict
harness-engineer
Provides guidance for a self‑improving engineering loop and relies on platform‑enforced sandboxing; no direct malicious behavior is defined.
⚠ 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)
References agent memory files
references/constraints.md · code · MEMORY.md
References agent configuration files
runtime/context-engineering.md · code · CLAUDE.md
Instructs covert action — may act without user awareness
references/sensitive-paths.md · prose · downgraded · silently
Popular HTTP library — network access
references/testing-standards.md · code · got
Permissions & capabilities
No declared permissions — minimal attack surface.
Is this flag fair?
Thanks — recorded.