claude-cleanup-audit
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/audit_claude_cleanup.pyexecutes local system utilities (scutil, dscl, systemsetup, security, launchctl, and ps) to gather configuration data for the audit report. - [PRIVILEGE_ESCALATION]: The instructions in
references/macos-identity.mddescribe usingsudoto modify system settings like the computer name and timezone. These operations are explicitly gated behind user confirmation in the skill's primary instructions. - [INDIRECT_PROMPT_INJECTION]: The skill reads and processes local JSON configuration files (
~/.claude.json,~/.claude/settings.json). While this creates a data ingestion surface, the skill uses structured JSON parsing and includes redaction logic to mitigate risks of data exposure in logs.
Audit Metadata