The Agent Skills Directory

[SAFE]: The skill enforces a rigorous approval process for all state-changing operations, including POST, PUT, PATCH, and DELETE requests, ensuring no administrative changes occur without user oversight. It also explicitly prohibits the agent from managing credentials or minting tokens directly.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes data from external third-party provider APIs and logs. Ingestion points: The agent reads external data using the call_platform_api and GET /log tools across all reference files. Boundary markers: The instructions lack specific delimiters or warnings to ignore instructions embedded within the ingested data. Capability inventory: The agent possesses write capabilities on the platform through approval-gated call_platform_api operations (e.g., in references/safe-admin-changes.md and references/create-sync-jobs.md). Sanitization: There are no explicit instructions for sanitizing or filtering external content before it is processed.

truto-operator