claude-permissions-audit
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a configuration management tool for Claude Code settings files (e.g., ~/.claude/settings.json). It identifies redundant entries and misaligned permissions to help users maintain a 'least privilege' and 'source of truth' model for their agent permissions.
- [SAFE]: No network activity or external data exfiltration patterns were detected. The skill operates exclusively on local configuration files and does not attempt to send data to any remote servers.
- [SAFE]: The skill does not contain any obfuscated content, hidden URLs, homoglyph attacks, or zero-width characters. All instructions are provided in plain, transparent markdown.
- [SAFE]: While the skill modifies security-relevant configuration files, it incorporates a clear multi-phase workflow that includes user review and explicit confirmation before applying any changes to the filesystem.
- [SAFE]: No remote code execution patterns, privilege escalation attempts, or persistence mechanisms were found. The skill utilizes standard file-system operations consistent with its stated purpose.
Audit Metadata