Security audit
Model Router
Security checks for vulnerabilities and agentic risk
Overview
This plugin transparently routes configured agents to different models and optionally narrows tools per turn, with no hidden persistence or unrelated data access found.
Before installing, configure it only for the agents you intend to route and keep the classifier allowedModels list tight, because prompts for configured agents are sent to the classifier model. Leave debug off unless troubleshooting, since debug logs can include prompt-derived text. Treat the README's optional host patch guidance as a separate manual change to review before applying.
SkillSpector was not run because this plugin release contains no bundled skills.
Endor dependency reachability
Checked Nov 9, 2026 · 5:18 PM UTC
No reachable vulnerabilities reported
Endor dependency scan completed.