subagent-driven-development

Pass

Audited by Gen Agent Trust Hub on May 29, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection via the processing of untrusted implementation plans.
  • Ingestion points: External implementation plan files (e.g., docs/plans/feature-plan.md) are read and extracted into task descriptions.
  • Boundary markers: Absent; the workflow does not include instructions to wrap external content in delimiters or add protective instructions when passing data to subagents.
  • Capability inventory: Dispatched subagents have high-privilege capabilities, including the ability to implement code, run tests, and perform git commits.
  • Sanitization: Absent; no validation or sanitization of the implementation plan content is performed before task extraction and subagent dispatch.
  • [NO_CODE]: The SKILL.md file contains only instructional text, process diagrams, and workflow descriptions. It does not include any executable scripts, binaries, or direct command-line instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
May 29, 2026, 04:27 PM
Security Audit — agent-trust-hub — subagent-driven-development