define-product
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed for structured documentation and interview workflows. Analysis of the instructions and provided templates reveals no malicious intent, obfuscation, or unauthorized access patterns.- [DATA_EXPOSURE_AND_EXFILTRATION]: The agent reads local documentation and code files (such as PRODUCT.md, GLOSSARY.md, and docs/decisions/README.md) to gather project context. This behavior is consistent with the skill's stated purpose and does not involve exfiltration to external domains or access to sensitive system credentials or environment variables.- [INDIRECT_PROMPT_INJECTION]: The skill ingests content from repository files, creating a potential attack surface for indirect prompt injection. This risk is effectively mitigated by the skill's design.
- Ingestion points: Reads from root PRODUCT.md, GLOSSARY.md, docs/decisions/README.md, and project code files.
- Boundary markers: Explicit instructions mandate keeping sources distinct (user statements vs. existing files) and requiring user confirmation for all changes to the product definition.
- Capability inventory: The skill is restricted to updating the single PRODUCT.md file at the repository root.
- Sanitization: The iterative interview process serves as a verification mechanism, ensuring the agent does not autonomously adopt instructions hidden within the data it processes.
Audit Metadata