offers
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: Analysis of the 10 provided files confirms the skill consists of instructional markdown content, worked examples, and evaluation criteria. No command execution, network activity, or obfuscation attempts were found.
- [INDIRECT_PROMPT_INJECTION]: The skill displays a surface for indirect prompt injection as it is designed to ingest context from external project files. 1. Ingestion points: Reads project-specific marketing context from '.agents/product-marketing.md', '.claude/product-marketing.md', or 'product-marketing-context.md'. 2. Boundary markers: Absent; the instructions do not prescribe delimiters or ignore-instructions for the ingested content. 3. Capability inventory: The skill does not define dangerous tools and relies on the agent's default processing capabilities. 4. Sanitization: Absent; no validation or filtering of the ingested file content is mentioned in the instructions.
Audit Metadata