skills/dachent/skills/ultraplan-codex/Gen Agent Trust Hub

ultraplan-codex

Pass

Audited by Gen Agent Trust Hub on Aug 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill enforces a strict read-only policy during the planning phase, explicitly prohibiting modifications to implementation files, dependency installation, or network operations.
  • [COMMAND_EXECUTION]: Local shell commands such as git status, git log, and rg (ripgrep) are used to gather repository context. These are standard operations for a developer assistant and are constrained to information gathering.
  • [SAFE]: The planning process includes a dedicated validation phase and an optional adversarial review where the plan is checked for correctness and risks before the user is asked to approve execution.
  • [SAFE]: All discovered context and planning state are stored locally in the .ultraplan/plan.md file, with no instructions or mechanisms identified for external data exfiltration.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 16, 2026, 07:58 AM
Security Audit — agent-trust-hub — ultraplan-codex