ClawAudit verdict
travel-swarm
Thin skill definition for a multi-MCP travel planning system integrating FlyAI, Gaode, Tencent, and Meituan APIs; no malicious behavior evident, API keys are standard configuration requirements.
⚠ 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 (9)
subprocess execution — runs system commands from Python
flyai_integration_report.md · code · subprocess.run(
subprocess with shell=True — command injection vector
flyai_integration_report.md · code · subprocess.run(cmd, shell=True
Possible hardcoded credential
backend/utils/food_nearby_api.py · prose · downgraded · API_KEY = "a8b1798825bfafb84c26bb5d76279cdc
HTTP request to bare IP address — common in malicious payloads
docs/COMPETITION_READY.md · code · http://47.253.101.130
Pipe to python — executes piped content as Python code
docs/COMPETITION_READY.md · prose · downgraded · | Python
<script> tag in markdown — potential code injection
main_7860.py · prose · downgraded · <script>
References agent memory files
GitHub开源标准教程_精细提示词.txt · code · MEMORY.md
Python os.getenv — reads environment variable
全真审计报告_TravelMaster_V6_20260411.txt · code · os.getenv(
Python aiohttp session — async network access
backend/mcp/amap_client.py · prose · downgraded · aiohttp.ClientSession
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.