blueprint
Pass
Audited by Gen Agent Trust Hub on Jul 2, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill consists exclusively of Markdown instructions (SKILL.md) and does not include any accompanying scripts, binaries, or automated installation processes.
- [SAFE]: No malicious patterns, such as prompt injection, credential harvesting, or unauthorized network communication, were detected in the instructions. The skill follows best practices for task decomposition and planning.
- [COMMAND_EXECUTION]: The skill directs the agent to use local development tools like Git and the GitHub CLI to verify project status. These commands are standard for software engineering tasks and are used solely to gather context for plan generation.
- [PROMPT_INJECTION]: There are no instructions intended to bypass safety guardrails or override the agent's core behavior. Instead, the skill includes a self-correcting 'adversarial review' step to identify errors or anti-patterns in the generated output.
Audit Metadata