executing-plans

Execute approved implementation plans in small user-checked batches.

139|7|Updated May 8, 2026
One-click install
npx skills add https://github.com/arbiterForge/codeArbiter --skill executing-plans-arbiterforge
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: executing-plans
Source: https://github.com/arbiterForge/codeArbiter/tree/main/plugins/ca/skills/executing-plans
Command: npx skills add https://github.com/arbiterForge/codeArbiter --skill executing-plans-arbiterforge

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents a model from freelancing through feature work by turning an approved plan into controlled, reviewable batches with human checkpoints.

Core Features & Use Cases

  • Batch Orchestration: Groups planned tasks into small execution batches so work stays ordered and manageable.
  • Delegated Implementation: Routes each batch to subagent-driven-development for fresh authoring, review, and verification instead of implementing inline.
  • Checkpoint Control: Stops after each batch for user acknowledgment, making it ideal for feature delivery that must remain auditable and gated.
  • Plan Validation: Refuses underspecified tasks, missing verification commands, and unresolved conflicts rather than guessing past them.

Quick Start

Use the executing-plans skill to carry out the approved plan for the current feature in small verified batches.

Frequently Asked Questions about executing-plans

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

FAQPage Schema
How do I stop Claude Code from improvising during plan execution?

To stop Claude Code from improvising during plan execution, you can use batch orchestration to group planned tasks into small, user-checked batches. This enforces fresh subagent execution and checkpoint pauses, keeping feature delivery auditable and gated.

How do I execute an approved implementation plan in verified batches?

Executing an approved implementation plan in verified batches involves routing grouped tasks to subagents for fresh authoring and review. The system stops after each batch for user acknowledgment, ensuring feature delivery remains controlled and auditable.

What happens to unresolved conflicts or underspecified tasks during feature delivery?

During feature delivery, unresolved conflicts or underspecified tasks trigger an escalation rather than the agent guessing past them. Plan validation actively refuses missing verification commands and unresolved conflicts to maintain strict execution boundaries.

Do I need a written plan and spec files to use batch orchestration?

Yes, you need a written plan, spec, context, and tech stack files to use batch orchestration. These files define task order and boundaries, and the workflow refuses underspecified tasks that lack the necessary details for fresh subagent execution.

How does subagent execution handle per-task verification and checkpointing?

Subagent execution handles per-task verification by routing each batch to fresh subagents for authoring, review, and verification. Checkpointing pauses work after every batch, requiring user acknowledgment before proceeding to the next set of planned tasks.