cc-build-loop
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill promotes security best practices by requiring the use of /security-review for sensitive code surfaces like authentication, payments, and user input.
- [INDIRECT_PROMPT_INJECTION]: The skill processes project-specific files such as docs/ROADMAP.md and docs/SECURITY-AUDIT.md to define its work. 1. Ingestion points: Reads task lists and plan files from the repository documentation. 2. Boundary markers: Explicitly requires the agent to restate success criteria and confirm scope with the user before starting implementation. 3. Capability inventory: File system writing, test suite execution, and internal tool usage. 4. Sanitization: No explicit sanitization of documentation content is defined.
- [COMMAND_EXECUTION]: The agent is instructed to run project test suites and platform review tools to verify implementation increments as part of the intended development loop.
Audit Metadata