/ac:execute

Executes approved Claude Code plans through wave-based orchestration and tier-routed workers.

3|Updated May 11, 2026

Search Tags:#automation #code review #plan execution #test verification #worker orchestration #auto mode #wave checkpoints

Ready to run

npx skills add https://github.com/anilcancakir/claude-code --skill ac-execute-anilcancakir

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: /ac:execute
Source: https://github.com/anilcancakir/claude-code/tree/main/plugins/ac/skills/execute
Command: npx skills add https://github.com/anilcancakir/claude-code --skill ac-execute-anilcancakir

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill turns an already-approved Claude Code plan into working changes across the repo without asking the user to approve each intermediate step.

Core Features & Use Cases

Wave-by-wave plan execution: Runs plan steps in dependency-aware waves until the plan is complete.
Tier-routed implementation with strict verification: Delegates steps to quick/junior/senior workers and enforces a four-layer verification flow per step.
Final gated deliverables: Performs a final code-review gate (standard or deep depending on plan complexity) and commits the result, optionally adding wave-after checkpoints for complex plans.

Quick Start

Run the executor for a prepared plan by issuing: /ac:execute <slug>.

Frequently Asked Questions about /ac:execute

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

FAQPage Schema

How do I automate execution of an approved development plan end-to-end?▼

Automated plan execution runs approved development steps to completion across your repository without requiring intermediate approvals. It applies wave-based orchestration, routing tasks to tiered workers and enforcing per-step verification before final code-review gating and commit generation.

What is wave-based orchestration for plan execution?▼

Wave-based orchestration executes plan steps in dependency-aware batches, running sequential waves until the plan is complete. Each step is delegated to tier-routed workers—quick, junior, or senior—and must pass a four-layer verification flow before the wave advances.

How do I run plan execution after generating a plan file?▼

Run plan execution by issuing the execute command with your plan slug. The executor processes the prepared plan file, running steps in dependency-aware waves, enforcing verification, performing final code review, and generating a commit with optional checkpoints for complex plans.

Can I run chained auto-mode execution for complex development plans?▼

Chained auto-mode execution is supported for complex plans. It runs steps continuously through wave-based checkpoints, applying tier-routed worker delegation and optional oracle parallelism, pausing only at wave-after checkpoints if the plan complexity requires them.

Does plan execution enforce code review before committing changes?▼

Plan execution enforces a final gated code review before committing. The gate applies either standard or deep review depending on the plan complexity, ensuring all prior step verifications and wave checkpoints have passed before the final commit and report are generated.

What are the limitations of automated plan execution without intermediate approval?▼

Automated plan execution requires a strictly approved plan file beforehand and cannot handle steps outside the prepared plan structure. It relies on tier-routed worker delegation and four-layer verification to prevent errors, meaning unstructured or loosely defined tasks may fail the verification gates.