feature-plan
Pass
Audited by Gen Agent Trust Hub on Apr 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No evidence of prompt injection, obfuscation, or malicious intent was found in the skill or its reference playbooks.
- [DATA_EXPOSURE]: The skill is designed to read project-level information (READMEs, design documents, logs, and screenshots) to provide context for planning. There are no network tools or data exfiltration patterns identified.
- [COMMAND_EXECUTION]: The instructions explicitly state that the skill is for planning and diagnosis only and must not enter code, script, or configuration implementation phases, effectively limiting the execution risk.
- [PROMPT_INJECTION]: The skill uses a structured 'Minimum Working Skeleton' and 'Problem Calibration' framework to maintain focus on user goals and prevent instruction drift or manipulation.
Audit Metadata