checkpoint-manager

Pass

Audited by Gen Agent Trust Hub on May 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a standard mechanism for managing session state checkpoints within a dedicated application directory (~/.ralph/checkpoints/).
  • [SAFE]: No sensitive data exposure or exfiltration patterns were identified. References to environment variables like ANTHROPIC_DEFAULT_*_MODEL are for configuration purposes.
  • [SAFE]: The skill does not contain any obfuscated code, remote downloads, or unauthorized command execution patterns.
  • [SAFE]: The instructions and allowed tools (Read, Write, Bash) are consistent with the stated purpose of managing local state files.
Audit Metadata
Risk Level
SAFE
Analyzed
May 3, 2026, 11:06 AM
Security Audit — agent-trust-hub — checkpoint-manager