memory-defrag
Pass
Audited by Gen Agent Trust Hub on Mar 2, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill consists entirely of markdown-based instructions for file organization and cleanup. It does not contain or reference any executable scripts, binaries, or automated code execution patterns.
- [DATA_EXFILTRATION]: All operations described are local to the agent's memory directory. There are no network requests, remote URLs, or mechanisms identified that would facilitate the unauthorized transmission of data to external servers.
- [PROMPT_INJECTION]: The instructions are focused on administrative file management and do not contain patterns designed to bypass safety filters, disclose system prompts, or override the agent's core behavioral constraints.
Audit Metadata