ClawAudit verdict
pansxng-websearch
While the skill auto-installs several dependencies on first use, this is transparently documented and serves the legitimate purpose of running a local SearXNG instance; no exfiltration or malicious behavior is present, though the broad auto-install behavior is worth user awareness.
⚠ 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)
Command substitution with curl — remote code execution
scripts/search.py · prose · downgraded · bash -c "$(curl
HTTP request to bare IP address — common in malicious payloads
scripts/search.py · prose · downgraded · http://127.0.0.1
subprocess execution — runs system commands from Python
scripts/search.py · prose · downgraded · subprocess.run(
Accesses shell history/config
scripts/search.py · prose · downgraded · ~/.zshrc
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class B/C/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.