Security audit
Model Requests
Security checks for vulnerabilities and agentic risk
Overview
This plugin provides the disclosed model-request monitoring page, including sensitive prompt and output excerpts, without evidence of external reporting or destructive behavior.
Install only if Gateway readers should see prompt excerpts, generated text, reasoning, and tool-call arguments across monitored sessions. Check configured provider URLs: without them, capture falls back to matching model endpoints on any host. Avoid secrets in URL query parameters. SkillSpector did not assess this plugin; this review inspected its backend and browser assets and checked interceptor behavior with synthetic requests, without testing a live Gateway.
SkillSpector was not run because this plugin release contains no bundled skills.
Static analysis
No suspicious patterns detected.