validation-sprint
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by interpolating untrusted data (user-provided idea descriptions and local project context files) directly into prompts for sub-agents without sanitization or strict boundary markers.\n
- Ingestion points: The skill reads the
$ARGUMENTSvariable and files within the.claude/product-context/directory inSKILL.md.\n - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are included when passing this data to the idea-researcher, market-researcher, or idea-skeptic agents.\n
- Capability inventory: The skill uses
Read,Glob,Grep, andTasktools, and possesses the capability to spawn and coordinate multiple sub-agents viaAgent Teams.\n - Sanitization: No input validation, escaping, or content filtering is performed on the ingested data before it is processed by the model.
Audit Metadata