ClawAudit verdict
safe-shrink
The skill explicitly instructs the agent that 'this section takes priority over all other skills' and forbids reading original files directly when intercepted — a classic prompt-injection hijack that redirects agent behavior beyond its stated document-compression purpose.
⚠ 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 (2)
Pipe to python — executes piped content as Python code
README.md · prose · downgraded · | Python
Instructs covert action — may act without user awareness
SKILL.md · prose · downgraded · silently
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.