product-marketing
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill scans local repository files (such as README, landing pages, and package.json) to automatically draft marketing context. This functionality introduces a surface where malicious instructions placed in those project files could influence the agent's behavior during the drafting phase.
- Ingestion points: Step 2 of
SKILL.mdidentifies ingestion of the README, landing pages, marketing copy, about pages, meta descriptions, and package.json. - Boundary markers: The skill relies on user verification ("The user then reviews, corrects, and fills gaps") to mitigate incorrect or malicious drafted content, although it does not define explicit text delimiters for the scanned data.
- Capability inventory: The skill performs file reads across the repository and writes results to
.agents/product-marketing.md. - Sanitization: The skill does not implement specific sanitization routines for the content ingested from the codebase.
Audit Metadata