subagent-driven-development

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns detected. The skill uses standard development tools and practices to manage task implementation and review.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection via the implementation plans it processes. 1. Ingestion points: Implementation plans (PLAN_FILE) are read by task-brief and review-package scripts. 2. Boundary markers: The implementer and reviewer prompts use Markdown headers to delimit task-specific requirements. 3. Capability inventory: The agent can execute git commands and write files within a dedicated workspace. 4. Sanitization: The skill relies on subagent prompt isolation and structured reporting to manage external data.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 01:43 PM
Security Audit — agent-trust-hub — subagent-driven-development