skills/razzeee/skills/commit/Gen Agent Trust Hub

commit

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses local git CLI commands for repository inspection and management.
  • [SAFE]: It explicitly directs the agent to ignore sensitive environment and configuration files during the staging process.
  • [SAFE]: Manual user confirmation is required before any commit or branch operations are finalized.
  • [PROMPT_INJECTION]: The skill processes external data from git history and file diffs (Ingestion points: git status, git diff, git log in SKILL.md; Boundary markers: absent; Capability inventory: git commit, git checkout, git add; Sanitization: absent), representing an indirect prompt injection surface that is mitigated by mandatory user review.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 11:23 PM
Security Audit — agent-trust-hub — commit