product-marketing
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted repository data to generate a centralized context document, creating an attack surface for indirect prompt injection and multi-step poisoning.
- Ingestion points: The agent scans files in the codebase including README, landing pages, and package.json to auto-draft marketing content.
- Boundary markers: The instructions do not define delimiters or warnings to ignore instructions embedded within the ingested repository files.
- Capability inventory: The skill possesses file read and write permissions to analyze the repo and save output to .agents/product-marketing.md.
- Sanitization: Content ingested from the codebase is summarized without explicit sanitization or filtering for malicious instructions.
- Multi-step impact: The generated document is intended as a foundational reference for all other marketing skills, potentially allowing injected instructions to influence downstream agent behavior.
Audit Metadata