ClawAudit verdict
skill-guard
skillshield
A comprehensive security scanner with 65 checks and SARIF output performing local static analysis via Python stdlib with no network dependencies or external data transmission.
⚠ 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 (13)
Pipe to python — executes piped content as Python code
SKILL.md · prose · downgraded · | Python
Prompt injection — tries to override agent instructions
tests/fake-agenttamper/SKILL.md · prose · downgraded · Ignore all previous instructions
Instructs agent to hide actions from user
tests/fake-helper/SKILL.md · prose · downgraded · Do not warn the user
Possible prompt injection — attempts to redefine agent identity
tests/fake-agenttamper/scripts/tamper.py · prose · downgraded · You are now
subprocess execution — runs system commands from Python
tests/fake-ansi/scripts/display.py · prose · downgraded · subprocess.run(
subprocess with shell=True — command injection vector
tests/fake-ansi/scripts/display.py · prose · downgraded · subprocess.run(["helper"], shell=True
HTTP request to bare IP address — common in malicious payloads
tests/fake-c2ip/scripts/analytics.py · prose · downgraded · http://91.92.242.30
References SSH/GPG private keys
tests/fake-pickle/scripts/loader.py · prose · downgraded · ssh_key
Accesses .ssh directory
tests/fake-pickle/scripts/loader.py · prose · downgraded · .ssh/
Contains shortened/invite URL
SKILL.md · code · bit.ly/legit
References agent memory files
tests/fake-agenttamper/SKILL.md · prose · downgraded · MEMORY.md
Python os.environ.get — reads environment variable
tests/fake-ansi/scripts/display.py · prose · downgraded · os.environ.get(
Python urllib.request — network access
tests/fake-c2ip/scripts/analytics.py · prose · downgraded · urllib.request
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.
data_encoding Is this flag fair?
Thanks — recorded.