Security audit
Skill Auto-Injection
Security checks for vulnerabilities and agentic risk
Overview
This plugin is purpose-aligned but needs review because it can automatically read prompts, scan local skills, send prompt text to a configurable model endpoint, and inject matched skill text into the agent context.
Install only if you are comfortable granting a community plugin access to conversation prompts and permission to modify the agent prompt. Keep the model endpoint local and trusted, remove broad trigger phrases or add an allowlist, avoid entering secrets while it is enabled, and check logs because skipped prompts may be recorded verbatim.
SkillSpector was not run because this plugin release contains no bundled skills.
Endor dependency reachability
Checked Nov 8, 2026 · 6:46 AM UTC
No reachable vulnerabilities reported
Endor dependency scan completed.
