task-breakdown

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill ingests untrusted external data in the form of architecture documents and feature specifications which may contain embedded instructions.
  • Ingestion points: SKILL.md specifies inputs as '[spec or architecture to decompose]', including architecture documents, feature specs, and problem descriptions.
  • Boundary markers: No specific boundary markers or 'ignore embedded instructions' warnings are present to delineate user-provided data from system instructions.
  • Capability inventory: Across all files, the only capability identified is writing output to '.agents/tasks.md'. The skill lacks network access, shell command execution, or sensitive file system operations.
  • Sanitization: There is no evidence of input sanitization or validation for the content of the provided specifications.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 08:14 AM