hyper-implement-loop

Executes a plan end-to-end by looping implement, code-review, and targeted fixes until blocking findings resolve.

3|Updated May 10, 2026
One-click install
npx skills add https://github.com/zeikar/hyperclaude --skill hyper-implement-loop
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hyper-implement-loop
Source: https://github.com/zeikar/hyperclaude/tree/main/skills/hyper-implement-loop
Command: npx skills add https://github.com/zeikar/hyperclaude --skill hyper-implement-loop

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

hyper-implement-loop removes the tedious back-and-forth of implementing a plan, getting Codex feedback, and iterating on fixes by running an autonomous implement → code-review → fix cycle until no blocking issues remain.

Core Features & Use Cases

  • End-to-end hardened implementation: Executes the plan to completion, then runs Codex code-review against the effective working tree and iteratively fixes blocking findings.
  • Critic-hardened loop with bounded reviews: Repeats review and targeted fixes until convergence or a strict maximum number of Codex reviews is reached.
  • Single-gesture workflow: Best for when you want the whole implement-hardening flow in one invocation rather than manual round-by-round control.
  • Use Case: Use it when you have a specific .hyperclaude/plans/*.md plan and want the repository updated with Codex gating on correctness, security, and build/test health.

Quick Start

Use hyper-implement-loop to execute a plan end-to-end by invoking: /hyperclaude:hyper-implement-loop /path/to/.hyperclaude/plans/your-plan.md

Frequently Asked Questions about hyper-implement-loop

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

FAQPage Schema
How do I automate code review and iterative fixes for repository changes?▼

You can automate code review and iterative fixes by running an implement → code-review → fix loop. This cycle executes your plan end-to-end and uses Codex to guard correctness, security, and build/test health until no blocking issues remain.

What is the best way to implement a plan with Codex code review gating?▼

The best way to implement a plan with Codex gating is using a single-gesture workflow that executes the plan and runs strict JSON Codex bridge reviews with semantic finding-map validation. It iteratively fixes blocking findings until convergence or a strict maximum number of reviews is reached.

Do I need experimental agent-teams support to run autonomous code fixes?▼

Yes, you need experimental agent-teams support to run autonomous code fixes. The workflow uses a live fixer teammate for iterative edits and requires it to execute the implement, code-review, and targeted fix cycles on your working tree.

Does Codex code review block style issues during the implementation loop?▼

Codex code review reports style issues as non-blocking during the implementation loop. It specifically guards correctness, security, and regressions as blocking findings, ensuring style problems do not halt the autonomous iteration cycle.

How do I execute a markdown plan file end-to-end with autonomous fixes?▼

To execute a markdown plan file end-to-end, invoke the implementation loop with the path to your `.hyperclaude/plans/*.md` file. It will apply repository code changes and iteratively patch the workspace until Codex reviews converge cleanly.

What happens if the Codex review loop does not converge on clean code?▼

If the Codex review loop does not converge, it stops after reaching a strict maximum number of Codex reviews. This bounded behavior prevents infinite cycles when blocking findings cannot be resolved by the live fixer teammate.