A professional workflow for shipping reviewer-ready production code through AI agents using strict contracts and the PACT protocol.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agentic-coding
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install agentic-coding using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Agentic Coding?
Agentic Coding is a high-discipline skill designed for developers who need merge-ready code from AI agents with explicit quality gates. Unlike general purpose assistants, this skill focuses on contract-first implementation and deterministic handoff checkpoints. By utilizing this within your Openclaw Skills collection, you ensure that every AI-generated change is backed by evidence and surgical precision.
The skill moves away from exploratory prompting toward a structured execution model. It is particularly effective for teams requiring high-reliability code, bug fixes with reproducible failures, and complex refactors. It serves as a bridge between raw AI output and professional engineering standards through rigorous validation and documentation.
Agentic Coding Use Cases
- Developing production-grade features with strict acceptance criteria
- Executing risky refactors requiring minimal side effects
- Resolving bugs with reproducible failure evidence
- Xcode-centered development including Swift features and iOS/macOS hotfixes
- Creating reviewer-ready handoff packets for complex pull requests
How Agentic Coding Works
- Lock a Contract: Define the objective, acceptance criteria, and non-goals before any code is written.
- Run the PACT Loop: Follow a four-step cycle of Problem framing, Acceptance design, Change set production, and Trace/Test evidence.
- Apply Surgical Diffs: Produce the smallest possible change set to meet the objective without unrelated code cleanup.
- Prove the Fix: For regressions, capture the failing state before applying and proving the resolution with evidence.
- Deliver Handoff: Generate a comprehensive packet containing changes, validation results, and rollback hints.
Agentic Coding Setup
Ensure you have git installed on your system. To initialize this component of your Openclaw Skills, create the local memory directory to enable persistence:
mkdir -p ~/agentic-coding/
touch ~/agentic-coding/memory.md
Upon first run, the skill will read setup.md to configure your environment. Any persistence remains explicitly opt-in to maintain your privacy.
Agentic Coding Data Schema & Taxonomy
The skill organizes its execution state within ~/agentic-coding/ using the following metadata taxonomy:
| File | Purpose |
|---|---|
| memory.md | Persistent developer preferences and operating modes. |
| contracts.md | Active task definitions and agreed-upon acceptance criteria. |
| evidence.md | Test snapshots and verification logs proving the fix. |
| handoffs.md | Delivery notes, blast radius analysis, and rollback hints. |
Agentic Coding Advanced Features
- Structured Escalation: Stops execution after two failed attempts to propose grounded alternatives rather than guessing.
- Xcode Integration: Direct support for tying acceptance to specific Apple targets, simulators, and test commands.
- PACT Protocol Enforcement: A deterministic loop that ensures problem framing and acceptance design precede code edits.
- Handoff-Grade Output: Automated generation of documentation that allows human reviewers to quickly validate AI contributions within the Openclaw Skills ecosystem.
SKILL.md
Loading
Related Openclaw Skills
Agent Identity & Personality
A framework for defining the identity, voice, and operational boundaries of AI agents to ensure authentic user interactions.
Agency
A comprehensive framework to build and operate a service-based agency including client management, project tracking, and automated pricing.
Affirmations
A personal mindset reinforcement system that builds and tracks custom daily affirmations.
Affiliate Marketing
A comprehensive toolkit for designing, managing, and scaling high-performance affiliate marketing programs with built-in fraud prevention and economic modeling.
Agentic Engineering
A tactical framework for maximizing productivity by orchestrating multiple AI coding agents through parallel workflows and pragmatic tool selection.
AGI / Artificial General Intelligence
A meta-cognitive reasoning layer that transforms AI agents from simple executors into human-level planners with self-awareness.
Featured*
