tooluniverse-self-review
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists entirely of instructional prompts designed to guide the model through a structured reasoning process (Recursive Expansion Tree). It does not contain any code, scripts, or commands.
- [SAFE]: The skill configuration explicitly includes
disable-model-invocation: true, which prevents the AI agent from calling any tools or performing network operations, effectively neutralizing most common attack vectors like data exfiltration or remote code execution. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted user data (the
taskandworkinputs). While this presents a surface for indirect prompt injection, the lack of tool access and external capabilities limits the potential impact to the quality of the generated text. No evidence of malicious instructions was found in the skill's own content.
Audit Metadata