clade-control

Start, stop, or clear blockers for Clade autonomous coding loops via a control endpoint.

8|5|Updated Feb 18, 2026

Search Tags:#claude code #developer productivity #autonomous coding #mcp server #loop control #api endpoint #remote orchestration

Ready to run

npx skills add https://github.com/shenxingy/Clade --skill clade-control

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: clade-control
Source: https://github.com/shenxingy/Clade/tree/main/adapters/openclaw/skills/clade-control
Command: npx skills add https://github.com/shenxingy/Clade --skill clade-control

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Remote users need a reliable way to control Clade’s autonomous coding loop without manually interacting with the server process.

Core Features & Use Cases

Start autonomous coding loops: Kick off a new loop for a specific project and goal, with configurable iteration, worker count, and model.
Stop loops gracefully: Send a stop request that allows the loop to finish its current task and exit cleanly.
Clear blockers: Remove detected blockers so a loop can be restarted and continue progressing.

Use cases: Start a loop to fix failing tests, stop an ongoing run when requirements change, or clear blockers when a previous run got stuck on an issue.

Quick Start

Ask your AI assistant to start a loop on the server by sending a start request with your project path, goal, and desired limits.

Frequently Asked Questions about clade-control

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

FAQPage Schema

How do I start an autonomous coding loop remotely for CI test fixes?▼

To start an autonomous coding loop remotely, send a POST request to the control endpoint with your project path, goal, and desired iteration limits. This initiates the Clade coding loop to automatically address failing tests or iterate on codebase improvements.

What is remote loop control for autonomous coding and how does it work?▼

Remote loop control allows you to manage autonomous coding loops via an API endpoint without manual server interaction. You send POST requests with JSON payloads to start, stop, or clear blockers for specific projects, receiving structured responses back from the server.

How do I stop a long-running coding loop gracefully when requirements change?▼

To stop a coding loop gracefully, send a stop request to the control endpoint. This allows the autonomous loop to finish its current task and exit cleanly, ensuring safe operational interruption of long-running runs without abrupt termination.

Do I need an API key to control remote coding loops?▼

Yes, you need to configure CLADE_BASE_URL and CLADE_API_KEY for authorization. These environment variables are required to authenticate your POST requests to the control endpoint when managing your autonomous coding loops.

How do I clear blockers on a stuck autonomous coding loop?▼

To clear blockers on a stuck autonomous coding loop, send a clear blocker request to the control endpoint. This removes detected issues so the loop can be restarted and continue progressing on its assigned goal.

Can I configure worker count and model selection when starting a coding loop?▼

Yes, when starting an autonomous coding loop, you can configure iteration limits, worker count, and model selection. These parameters are included in the JSON payload sent to the control endpoint to customize the loop execution.