principle-build-the-lever
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a purely instructional document outlining software engineering best practices for AI agents. It does not contain any executable code, network operations, or sensitive data access.
- [DYNAMIC_EXECUTION]: The skill's primary methodology is instructing the agent to generate and execute code (e.g., codemods, generators, and scripts) to perform tasks. While this promotes the use of dynamic execution, it is framed as a controlled development workflow where the generated code is an artifact for human review. It does not provide any pre-written malicious scripts or remote execution commands.
- [INDIRECT_PROMPT_INJECTION]: The skill provides guidance on delegating tasks to child agents via shared instruction sets. It explicitly includes security-conscious advice to keep these instructions outside the write-scope of the child agents to prevent them from modifying their own 'contracts' or guidelines.
Audit Metadata