ClawAudit verdict
videomp3word-mcp
MCP server that sends media URLs to a configured upstream transcription service for structured knowledge extraction; the skill explicitly documents data sharing with upstream endpoints and recommends auditing them before deployment.
⚠ 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 (7)
Possible hardcoded credential
SKILL.md · frontmatter · API_KEY: "Optional. Upstream API key sent alongside the session cookie.
<script> tag in markdown — potential code injection
src/ui/playground.ts · prose · downgraded · <script>
References child_process — can spawn system processes
src/smoke.test.ts · prose · downgraded · child_process
Dynamic import() — loads module at runtime
src/smoke.test.ts · prose · downgraded · import("
HTTP request to bare IP address — common in malicious payloads
src/smoke.test.ts · prose · downgraded · http://127.0.0.1
Python urllib.request — network access
sdk/python/videomp3word_client.py · prose · downgraded · urllib.request
Accesses sensitive environment variables
src/config.ts · prose · downgraded · process.env.MCP_ACCESS_KEY
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class B/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.