salla-docs

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The workflow fetches and ingests readable text from public Salla documentation pages at runtime (e.g., https://docs.salla.dev/<id>.md and https://docs.salla.dev/llms.txt) and then reads the inline OpenAPI YAML blocks into the agent’s LLM context, which is outsider-authored free text from a third-party website.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent to fetch live pages from https://docs.salla.dev/.md (and to use https://docs.salla.dev/llms.txt to locate pages) at runtime and to read the inline OpenAPI blocks as the authoritative schema for validation and upgrade logic, so external content directly controls the agent's behavior.