hierarchical-reasoning

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWPROMPT_INJECTION
Full Analysis
  • PROMPT_INJECTION (LOW): The skill possesses a surface for indirect prompt injection within its reasoning pipeline. \n
  • Ingestion points: The skill accepts untrusted data through the problem description argument and the --context JSON flag in the scripts/hierarchical_reasoner.py invocation. \n
  • Boundary markers: Absent. The documentation does not specify the use of delimiters or 'ignore' instructions to isolate user-provided content from the reasoning framework. \n
  • Capability inventory: Capabilities are limited to generating internal reasoning states and synthesis. No file writing, network operations, or arbitrary command execution capabilities were found in the provided files. \n
  • Sanitization: No input validation, escaping, or filtering of the problem description is identified in the provided documentation.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 05:02 AM