verification-before-completion
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is entirely instructional and establishes a defensive practice requiring empirical evidence (such as test logs and build exit codes) before an agent reports success.
- [NO_CODE]: No executable scripts, binaries, or automated tasks are included. The skill only contains natural language instructions and visual examples.
- [PROMPT_INJECTION]: The instructions do not contain patterns intended to bypass safety filters, extract system prompts, or disregard safety guidelines. While it uses imperative language to enforce its verification logic, the intent is task-oriented and safety-positive.
Audit Metadata