paywalls
Pass
Audited by Gen Agent Trust Hub on Jul 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or high-risk behaviors were detected. The skill is entirely documentation-based and does not execute code or access sensitive system resources.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it is designed to ingest data from local project files to provide context.
- Ingestion points: .agents/product-marketing.md, .claude/product-marketing.md, and product-marketing-context.md as referenced in SKILL.md.
- Boundary markers: None present in the instructions.
- Capability inventory: None. The skill does not request or utilize any tools for subprocess execution, network operations, or file writing.
- Sanitization: None present.
- Risk Evaluation: Because the skill lacks any dangerous capabilities, the risk associated with reading these context files is negligible.
Audit Metadata