lean-build
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from user requests and repositories to derive implementation goals. If this data contains malicious instructions, the agent might follow them during the development process.
- Ingestion points: Processes 'request' and 'repository' content (SKILL.md).
- Boundary markers: No specific boundary markers or instructions to ignore embedded content are provided.
- Capability inventory: The skill involves implementation tasks which typically require file writing and code execution capabilities.
- Sanitization: No sanitization or validation of the input request or repository content is specified.
Audit Metadata