ad-copy-compliance-review
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill operates as described, providing a utility for ad compliance review. Analysis of the provided scripts and configuration confirms that network operations are restricted to the vendor's infrastructure at ai-skills.ai. Sensitive credentials are handled via environment variables (AISKILLS_API_KEY), adhering to standard security practices. No evidence of command injection, persistence mechanisms, or privilege escalation was found.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external data, which inherently creates a surface for indirect prompt injection. However, this is consistent with its primary purpose.
- Ingestion points: Untrusted content enters the system via the materialUrl, materialFile, and materialText parameters as defined in SKILL.md and references/form-schema.json.
- Boundary markers: The provided files do not specify explicit delimiters or instructions to the model to ignore embedded commands within the processed materials.
- Capability inventory: The skill's capabilities involve processing text to generate summaries and advice; the scripts/run.py tool is limited to standard HTTP POST and GET requests to the vendor's API.
- Sanitization: No specific filtering or sanitization logic for user-provided materials was identified in the static configuration. Given the intended use case and lack of dangerous tool access, this surface is considered a standard functional characteristic of the skill.
Audit Metadata