Prompt-injection style instruction pattern detected.
Warn
- Code
- suspicious.prompt_injection_instructions
- Location
- skills/ai-guardian/references/agent-guardrails.md:56
- Evidence
Copy this into your agent's system prompt:
Security audit
AI Guardian
Security checks for vulnerabilities and agentic risk
Overview
AI Guardian is a disclosed local-LLM observability and governance plugin with sensitive but purpose-aligned model-management abilities.
Install this only if you want an agent to inspect and govern local LLM runtimes. Be aware that the full MCP server includes write tools that can pull, unload, or remove local models and change allow/deny policy; use a restricted account or expose only read/scan tools for observe-only use. Optional endpoint tokens are documented as encrypted, and prompt logs are documented as redacted rather than storing raw prompts.
SkillSpector was not run because this plugin release contains no bundled skills.
Static analysis
Detected: suspicious.prompt_injection_instructions