ha-integration-knowledge
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to process external files, which creates a potential surface for indirect prompt injection.
- Ingestion points: Integration code, tests, and manifest files located in the
./homeassistant/components/directory. - Boundary markers: The instructions lack specific boundary delimiters or guidelines to help the agent ignore instructions potentially embedded in the data being analyzed.
- Capability inventory: The skill facilitates the agent's use of file-reading and testing capabilities on local source code files.
- Sanitization: No procedures for sanitizing or validating the contents of the integration files are provided.
Audit Metadata