answer-ads-questions
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources, specifically the Meta Marketing API and local markdown files in the ads/ directory. This creates a surface for potential indirect prompt injection attacks.
- Ingestion points: Data is pulled into the agent's context from Meta's API responses and local campaign documentation (state.md, decisions.md, README.md).
- Boundary markers: The skill does not define specific delimiters or instructional warnings to prevent the agent from following directions that might be embedded in the ad metadata or documentation.
- Capability inventory: Across all adapter modes, the skill's capabilities are restricted to read-only API calls and local file access; it lacks the ability to execute code, write files, or perform unauthorized network operations.
- Sanitization: No explicit sanitization or filtering of the ingested external content is described.
- [DATA_EXFILTRATION]: The skill uses network requests to interact with the Meta Graph API for legitimate data retrieval. The documentation includes strong security guidance to ensure that API tokens are handled securely within the environment and are never exposed in user interactions or log files.
Audit Metadata