om-auto-fix-pr

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill incorporates robust defensive instructions in references/agentic-setup.md to ignore malicious directives embedded in untrusted repository content, such as PR bodies or commit messages, specifically mentioning bypasses like 'ignore previous instructions'.
  • [COMMAND_EXECUTION]: Uses standard shell commands for managing Git worktrees and executing repository validation scripts. These operations are essential for its purpose and are scoped to the project environment.
  • [DATA_EXFILTRATION]: Implements mandatory 'Secrets hygiene' rules in references/rules.md that prevent the leakage of environment variables, tokens, or credentials into public PR comments or logs, even if instructed otherwise by external input.
  • [SAFE]: The skill follows established patterns for agentic development tools, using repository-defined configuration for its operational parameters and maintaining clear boundaries between untrusted data and executable instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 04:31 AM
Security Audit — agent-trust-hub — om-auto-fix-pr