Caveat verdict
llm-proxy
Local LLM API proxy with content safety auditing (dangerous command detection, PII/credential filtering) that listens only on 127.0.0.1:18888; acts as a security layer rather than a threat, with no exfiltration patterns.
⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.
Automated static analysis — not a human review. Caveat flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.
Permission integrity
network_out
Findings (8)
HTTP request to bare IP address — common in malicious payloads
SKILL.md · code · http://127.0.0.1
Possible hardcoded credential
scripts/content-filter-rules.json · prose · downgraded · password=",
Pipe-to-python pattern — remote code execution risk
scripts/llm-proxy-ctl.sh · prose · downgraded · curl -s "$PROXY_URL" | python
Pipe to python — executes piped content as Python code
scripts/llm-proxy-ctl.sh · prose · downgraded · | python3
References sudo — requests elevated privileges
scripts/content-filter-rules.json · prose · downgraded · sudo
Sets world-executable permissions
scripts/llm-proxy-config.json · prose · downgraded · chmod 777
Python urllib.request — network access
scripts/llm-proxy.py · prose · downgraded · urllib.request
Python os.environ.get — reads environment variable
scripts/llm-proxy.py · prose · downgraded · os.environ.get(
Permissions & capabilities
No declared permissions — minimal attack surface.
network_out Is this flag fair?
Thanks — recorded.