product-marketing
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's 'Auto-draft from codebase' feature reads various project files (README, package.json, marketing copy) and incorporates their content into a central marketing document. If an attacker can control the content of these local files, they could inject instructions that might be executed by this agent or other agents reading the resulting context file.
- Ingestion points: Local project files such as README, landing pages, and package.json (SKILL.md).
- Boundary markers: The generated markdown template does not use specific markers to delineate or neutralize instructions potentially found in the imported text.
- Capability inventory: Includes file system read access for repository analysis and write access for document maintenance (SKILL.md).
- Sanitization: There is no evidence of sanitization or safety checks performed on the text extracted from the analyzed files.
Audit Metadata