subagent-driven-development
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill orchestrates a development process that ingests content from implementation plans and interpolates it into subagent instructions.
- Ingestion points: Plan files (e.g.,
docs/plans/feature-plan.md) are read, and their task descriptions are pasted directly into the prompts for implementer and reviewer subagents as seen inimplementer-prompt.mdandspec-reviewer-prompt.md. - Boundary markers: The prompts utilize markdown headers (e.g.,
## Task Description,## What Was Requested) to separate task content, but lack explicit instructions for the subagent to disregard control sequences or commands within the ingested data. - Capability inventory: The subagents spawned by this workflow have the capability to write code, execute test suites, and perform git operations.
- Sanitization: The skill does not implement validation or sanitization of the extracted task text before it is provided to the subagents.
Audit Metadata