oac-approach

Proposes a concise implementation plan for user approval before writing any code.

4.8k|390|Updated Aug 14, 2025
One-click install
npx skills add https://github.com/darrenhinde/OpenAgentsControl --skill oac-approach
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: oac-approach
Source: https://github.com/darrenhinde/OpenAgentsControl/tree/main/plugins/claude-code/skills/oac-approach
Command: npx skills add https://github.com/darrenhinde/OpenAgentsControl --skill oac-approach

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

AI coding agents often jump straight into writing code without understanding project conventions, producing generic output that requires heavy rework. This Skill enforces a plan-first workflow so every implementation starts from shared understanding and explicit approval.

Core Features & Use Cases

  • Context Discovery: Invokes oac:context-discovery to load project standards and patterns before proposing an approach.
  • Approval Gate: Blocks all code changes until the user explicitly approves a lightweight proposal covering what, how, assumptions, and affected files.
  • Smart Handoff: Routes simple tasks (1-3 files) directly to implementation and complex work (4+ files) to oac:task-breakdown for subtask planning.
  • Use Case: A developer asks the agent to add a new API endpoint. Instead of immediately generating code, the agent discovers the project's validation and ORM patterns, proposes a short plan referencing those standards, and waits for approval before implementing.

Quick Start

Ask the agent to implement a feature and it will propose an approach for your approval before writing any code.

Frequently Asked Questions about oac-approach

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

FAQPage Schema
How do I make an AI coding agent plan before writing code?

Use a plan-first workflow skill that blocks code changes until you approve a proposed approach. The agent reads your request, discovers project context, presents a short proposal with assumptions and affected files, and waits for explicit approval.

How to get AI-generated code to follow my project conventions?

Run context discovery before implementation so the agent loads your project's standards and pattern files. The proposal then references those conventions, and the resulting code matches your validation, ORM, and response formats instead of generic output.

What happens if no project context files are installed?

The agent proceeds anyway, noting 'none' in the proposal and using general best practices. It appends a tip suggesting you run /install-context to set up context files for better convention-aligned results.

When does the agent ask clarifying questions before proposing?

Questions are asked only when the request is genuinely ambiguous or context discovery reveals missing details like framework or database choice. Otherwise the agent states assumptions directly in the proposal rather than interrogating you.

How are complex features handled after plan approval?

Simple tasks touching 1-3 files are implemented directly after approval. Complex work spanning 4 or more files is handed off to oac:task-breakdown, which creates subtasks before implementation begins.