memory-merger
Pass
Audited by Gen Agent Trust Hub on Feb 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's file access is restricted to directories intended for agent configuration, such as
vscode-userdata:/User/prompts/and.github/instructions/. It does not interact with sensitive system files or private user credentials. - [SAFE]: A mandatory human-in-the-loop checkpoint is integrated into the workflow. The agent must present proposed changes and wait for user approval (the 'go' command) before any files are created, updated, or modified, effectively mitigating risks from processing untrusted data.
Audit Metadata