subagent-driven-development
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: No security issues were detected. The skill consists of process documentation and instructions for a 'Subagent-Driven Development' workflow. It does not contain executable scripts, configuration files that modify the system, or requests for sensitive data.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a process for ingesting external data, which creates a potential surface for indirect prompt injection.
- Ingestion points: Implementation plans are loaded from
docs/plans/and used to generate instructions for subagents. - Boundary markers: The workflow uses a structured markdown template in Step 2 to delimit task instructions, context, and success criteria.
- Capability inventory: The workflow assumes capabilities like file modification and test execution will be available to the subagents.
- Sanitization: No specific sanitization or filtering of the plan files is described, though the 'Review Changes' step (Step 3) serves as a manual human-in-the-loop validation for any generated output.
Audit Metadata