task-decomposer

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its primary input mechanism.
  • Ingestion points: Untrusted data enters the agent's context through the 'problem-description' argument defined in SKILL.md.
  • Boundary markers: The instructions lack explicit delimiters or markers to isolate the user-provided task description from the agent's instructions.
  • Capability inventory: The skill is permitted to use file-system tools (Read, Grep, Glob) which could be leveraged to access sensitive files if the agent is manipulated by injected instructions in the description.
  • Sanitization: There is no logic provided to sanitize or validate the user-provided input before processing.
  • [NO_CODE]: The skill is composed entirely of instructional content and metadata; no executable code, Python scripts, or Node.js packages are included.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 06:13 PM