execute-review

Review one completed execution phase against a plan with Claude-Codex.

1|Updated Apr 6, 2026
One-click install
npx skills add https://github.com/Honigbart/honeyflow --skill execute-review
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: execute-review
Source: https://github.com/Honigbart/honeyflow/tree/main/execute-review
Command: npx skills add https://github.com/Honigbart/honeyflow --skill execute-review

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill guides the safe, bounded review of exactly one completed execution phase by running a Claude-Codex loop, ensuring traceable state and archive updates for each phase.

Core Features & Use Cases

  • Bound one-phase Claude-Codex review flow that can resume an unfinished phase.
  • Automatically identifies the earliest eligible done phase and initializes per-phase review tracking.
  • Archives phase reviews and updates plan state, enabling end-to-end delivery governance.

Quick Start

Invoke the execute-review skill to review exactly one completed phase and update the plan state accordingly.

Frequently Asked Questions about execute-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate plan execution review for completed phases?

To review one completed execution phase, invoke the skill to run a bounded Claude-Codex review loop. It automatically identifies the earliest eligible done phase, initializes per-phase review tracking, archives the review, and updates the plan state accordingly.

What files do I need to resume a phase review?

You need the plan's final_plan.md and execution_state.md files to exist. These artifacts provide the required plan context and execution state to safely resume the review and touch only phase-specific artifacts.

Does the Claude-Codex review loop work with Ollama?

Yes, the Claude-Codex review loop supports optional Ollama-assisted review. This integration helps examine exactly one completed execution phase against a plan while maintaining traceable state and archive updates.

How does per-phase status tracking work during execution?

Per-phase status tracking works by automatically identifying the earliest eligible done phase and initializing review metadata for it. This ensures traceable state updates and archiving for each phase, enabling end-to-end delivery governance across the plan.

What is the best way to bound an automated code review to a single phase?

Bounding a review to a single phase requires a bounded Claude-Codex loop that touches only phase-specific artifacts. This approach ensures the review examines exactly one completed execution phase against the plan, preventing scope creep into other phases.