thought-based-reasoning

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides structured reasoning techniques such as Chain-of-Thought, Tree of Thoughts, and ReAct to assist the agent in solving complex tasks. These are standard prompt engineering practices.
  • [METADATA_POISONING]: Minor inconsistencies were noted between the frontmatter (version 1.0.0, author ClawFu) and the footer (version 2.0, author GUIA). These appear to be unintentional documentation discrepancies and do not impact security.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides templates for processing external problem statements.
  • Ingestion points: Problem statements processed in CoT, ToT, and PAL templates.
  • Boundary markers: The templates use markdown headers and structured sections to delineate reasoning from inputs.
  • Capability inventory: The skill itself lacks executable code or network access; it suggests utilizing the agent's existing tools (e.g., Python for PAL).
  • Sanitization: No explicit sanitization or input validation is performed on the problem statements.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 08:26 PM
Security Audit — agent-trust-hub — thought-based-reasoning