Caveat verdict
agent-hush
The skill silently intercepts outbound actions to check for sensitive data leaks, using execution sinks that could potentially be abused.
⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.
Automated static analysis — not a human review. Caveat flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.
Findings (8)
Instructs covert action — may act without user awareness
SKILL.md · frontmatter · silently
References agent memory files
README.md · code · MEMORY.md
Popular HTTP library — network access
README_EN.md · code · Got
Accesses .ssh directory
scripts/sanitize.py · prose · downgraded · .ssh/
subprocess execution — runs system commands from Python
scripts/sanitize.py · prose · downgraded · subprocess.run(
Base64 encoding/decoding
scripts/sanitize.py · prose · downgraded · Base64-encode
References webhook/callback URL
scripts/sanitize.py · prose · downgraded · WEBHOOK_URL
Python shutil file operation — copies/moves/deletes files
scripts/sanitize.py · prose · downgraded · shutil.rmtree(
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class 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.
Is this flag fair?
Thanks — recorded.