feature-dev
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external codebases throughout its lifecycle, including exploration, design, and testing phases.
- Ingestion points: User-provided input is ingested via
$ARGUMENTSinSKILL.md. Additionally, the skill reads files from the project codebase in Phase 2 (via exploration agents), Phase 5 (implementation), and Phase 6 (test generation). - Boundary markers: The instructions do not define clear boundaries or provide warnings to the agent to disregard instructions that might be embedded within the files being analyzed.
- Capability inventory: The skill is capable of reading files, launching sub-agents, writing code to the local filesystem, and updating task lists via
TodoWrite. - Sanitization: There is no evidence of content sanitization, filtering, or validation for the data retrieved from the codebase files before it is processed by the agent.
Audit Metadata