principle-build-the-lever
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of instructional content for developers or agents. It defines a philosophy for task automation and verification that encourages building rerunnable artifacts like scripts or codemods instead of manual edits.\n- [NO_CODE]: There are no scripts, binaries, or executable commands contained within this skill. The frontmatter includes
disable-model-invocation: true, which further restricts the agent's behavior when this skill is active.\n- [PROMPT_INJECTION]: No evidence of prompt injection, jailbreak attempts, or safety bypass instructions were found. The language is purely methodological.\n- [EXTERNAL_DOWNLOADS]: No external URLs or remote resource fetches are present. The internal links provided refer to other local documentation files within the same skill set.
Audit Metadata