pb-subagent-driven-development
Pass
Audited by Gen Agent Trust Hub on Jun 22, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill processes implementation plans and task briefs that are interpolated into subagent prompts. This creates a surface for indirect prompt injection if the plan files contain untrusted content. However, the skill implements isolation through fresh subagent contexts per task and requires explicit verification by a reviewer subagent, which aligns with secure development practices.
- [COMMAND_EXECUTION]: The skill includes several bash utility scripts (review-package, sdd-workspace, task-brief) used for repository maintenance and workflow orchestration. These scripts perform standard file operations and git commands within the project's working directory and do not execute remote code or handle untrusted input unsafely.
Audit Metadata