power-of-ten
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a set of rigorous coding standards based on industry best practices for safety-critical software. The instructions guide the agent to improve code safety, predictability, and reliability.
- [COMMAND_EXECUTION]: The instructions require the agent to "Run all ten inspections," "Verify affected behavior," and "Record the exact commands, settings, and results." While this involves shell command execution, it is restricted to running diagnostics, static analysis tools, and tests on the code the agent is tasked with modifying. This is standard behavior for a development-focused AI agent.
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process and analyze external source code. This represents a theoretical surface for indirect prompt injection (e.g., instructions hidden in code comments), but the skill's requirement for structured evidence and verification of behavior serves as a functional control. The risk is consistent with standard code-processing tasks.
Audit Metadata