home-assistant-dev

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt includes config flows that collect passwords and an example diagnostics function that returns dict(entry.data) (which may contain credentials), forcing the LLM to handle and potentially output secret values verbatim.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The docs include explicit runtime commands that pull and run remote container images (executing their code) such as ghcr.io/home-assistant/amd64-builder and other GHCR base images (e.g., ghcr.io/home-assistant/armhf-base:latest), which are external runtime dependencies that execute remote code.