dispatching-parallel-agents

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWNO_CODE
Full Analysis
  • NO_CODE (SAFE): The skill consists of a single Markdown file providing instructional patterns. No scripts, binaries, or automated tools are included.
  • INDIRECT_PROMPT_INJECTION (LOW): The documented pattern involves agents reading external data (test logs) to perform code modifications, which is a known vulnerability surface. However, the skill does not provide the mechanism for this, only the conceptual framework.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 04:22 AM