subagent-driven-development
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill identifies an attack surface for indirect prompt injection as it processes externally defined implementation plans.
- Ingestion points: The skill reads task descriptions from implementation plans (e.g.,
docs/superpowers/plans/feature-plan.md). - Boundary markers: Task content is extracted and interpolated into subagent prompts without explicit delimiters or warnings to ignore embedded instructions.
- Capability inventory: Subagents are granted full development capabilities, including file system access, shell execution for testing, and git operations.
- Sanitization: No sanitization or validation of plan data is performed before it is passed to subagents.
- [COMMAND_EXECUTION]: The skill uses subagents to execute code and tests as part of the intended development workflow. It includes safety constraints such as requiring execution on non-protected branches and utilizing isolated git worktrees.
Audit Metadata