catch-up
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides instructions for the model to recap the conversation status using a specific template. It does not contain any executable code or scripts.
- [SAFE]: The skill configuration explicitly disables model invocation of other tools (
disable-model-invocation: true), preventing any potential for secondary tool-based attacks. - [SAFE]: No patterns of prompt injection, data exfiltration, or obfuscation were identified. The logic for recommending a different skill (
checkpoint) in specific contexts is a benign workflow instruction.
Audit Metadata