ClawAudit verdict
xinyi-drink
The skill seems to be focused on providing information about drinks and menus, with no apparent malicious intent.
⚠ 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 (8)
Dynamic __import__('os') — Python OS command execution
spec/install_test.py · prose · downgraded · __import__("os")
Writes to SKILL.md — self-modifying skill
spec/install_test.py · prose · downgraded · SKILL.md").write
HTTP request to bare IP address — common in malicious payloads
spec/claim_reward_test.py · prose · downgraded · http://127.0.0.1
subprocess execution — runs system commands from Python
spec/install_test.py · prose · downgraded · subprocess.run(
Python os.getenv — reads environment variable
scripts/skill_config.py · prose · downgraded · os.getenv(
Python urllib.request — network access
scripts/skill_http.py · prose · downgraded · urllib.request
Python shutil file operation — copies/moves/deletes files
spec/install_test.py · prose · downgraded · shutil.copytree(
Python os.environ.get — reads environment variable
spec/skill_config_test.py · prose · downgraded · os.environ.get(
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.