task-dependency-patterns

Pass

Audited by Gen Agent Trust Hub on Feb 20, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [Indirect Prompt Injection] (LOW): The task management patterns involve processing and acting upon user-supplied task subjects and descriptions, which serves as an ingestion surface for indirect prompt injection.\n
  • Ingestion points: The scripts/task-tree-visualizer.py utility processes task JSON from stdin, and agents retrieve task data using TaskList and TaskGet tools.\n
  • Boundary markers: No specific delimiters or instructions to ignore nested directives within task metadata are established in the rules or reference materials.\n
  • Capability inventory: The skill provides a Python script for execution and defines patterns for spawning subagents and teammates via the Task tool.\n
  • Sanitization: No input validation, escaping, or sanitization logic is implemented for the task content fields.\n- [Static Analysis] (SAFE): The documentation and rules strictly adhere to its stated purpose of improving workflow efficiency and multi-agent coordination. No prompt injection overrides or obfuscation techniques were found.\n- [Code Analysis] (SAFE): The scripts/task-tree-visualizer.py script is a secure implementation that uses only standard Python libraries and includes cycle detection to prevent infinite recursion when visualizing dependency trees.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 20, 2026, 10:05 PM