Caveat verdict
autonomous-procurement-agent
Enterprise procurement quote parsing tool that explicitly documents its Privacy Shield scrubbing sensitive fields before any external API call, runs locally without OPENAI_API_KEY, and uses webhook for license activation; all behavior matches its fraud-detection stated purpose.
⚠ 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
package_install
Findings (11)
Possible hardcoded credential
SKILL.md · code · SECRET="your_ls_webhook_secret
Pipe to python — executes piped content as Python code
SKILL.md · code · | python3
Recursive delete from root or home — destructive command
PRIVACY.md · code · rm -rf ~
Instructs covert action — may act without user awareness
self-healing-parser.js · prose · downgraded · silently
Uses exec() — may execute shell commands
self-healing-parser.js · prose · downgraded · exec(
Popular HTTP library — network access
SKILL.md · prose · downgraded · got
References tunneling service
SKILL.md · prose · downgraded · ngrok
References webhook/callback URL
auth-middleware.js · prose · downgraded · WEBHOOK_URL
Node http/https module — low-level network access
auth-middleware.js · prose · downgraded · require('https')
Base64 encoding/decoding
auth-middleware.js · prose · downgraded · base64-encode
Accesses sensitive environment variables
self-healing-parser.js · prose · downgraded · process.env.OPENAI_API_KEY
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class A). 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.
network_inpackage_install Is this flag fair?
Thanks — recorded.