A comprehensive AI agent skill for Windsurf IDE that enables autonomous coding, multi-file editing, and complex terminal-based workflows.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install windsurf-cascade
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 windsurf-cascade 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 Windsurf Cascade Agent Skill?
The Windsurf Cascade Agent Skill provides a robust framework for interacting with the Cascade AI within the Windsurf IDE. Unlike traditional autocompletion tools, Cascade acts as a flow-aware agentic assistant that understands the entire codebase, tracks real-time developer actions, and executes complex tasks autonomously. By leveraging the principles of Openclaw Skills, this integration allows developers to create modular, reusable sets of instructions and checklists that the AI can invoke to maintain high standards of code quality and architectural consistency.
At its core, this skill bridges the gap between simple chat interfaces and full-scale development orchestration. It supports multiple frontier models including SWE-1.5, Claude 4.6, and GPT-5.2, providing the flexibility to choose the right intelligence level for specific tasks. Whether you are refactoring legacy modules or deploying new microservices, this skill ensures that your AI agent remains aligned with your project rules and historical context through sophisticated memory management and dedicated terminal integration.
Windsurf Cascade Agent Skill Use Cases
- Automating cross-file refactoring and architectural updates with high precision.
- Standardizing PR reviews by automating branch checkouts and addressing feedback via CLI integration.
- Orchestrating complex deployment sequences using custom Openclaw Skills checklists and workflows.
- enforcing project-specific coding standards through automated workspace rules and memories.
- Accelerating bug resolution by allowing the agent to analyze error logs and apply fixes directly in the terminal.
How Windsurf Cascade Agent Skill Works
- Contextual Ingestion: Cascade continuously monitors your workspace, terminal output, and active edits to maintain a synchronized state with your development flow.
- Interaction Mode Selection: Users initiate either Write Mode for active codebase modifications or Chat Mode for read-only research and analysis.
- Skill Discovery: The agent scans the
.windsurf/skills/directory to identify relevant Openclaw Skills that match the current task's requirements. - Tool Execution: Cascade utilizes built-in tools for semantic search, web browsing, and terminal execution, or calls external services via the Model Context Protocol (MCP).
- Verification and Reversion: All agent actions are recorded via named checkpoints, allowing developers to review diffs and revert changes if the trajectory deviates from the desired outcome.
Windsurf Cascade Agent Skill Setup
To get started with this skill, download the Windsurf Editor and install the CLI tools for seamless terminal integration.
# Install the windsurf command for macOS
sudo ln -sF /Applications/Windsurf.app/Contents/Resources/app/bin/windsurf /usr/local/bin/windsurf
# Open your project to initialize context
windsurf ./my-project
After installation, configure your environment by importing settings from VS Code or Cursor. Create a .windsurf/skills directory in your project root to begin defining your first Openclaw Skills to guide the agent's behavior.
Windsurf Cascade Agent Skill Data Schema & Taxonomy
The skill manages data across several specialized directories and configuration files within your workspace to ensure logical organization.
| Directory/File | Description | Usage |
|---|---|---|
.windsurf/skills/ |
Modular instruction sets | Stores SKILL.md and supporting templates for Openclaw Skills. |
.windsurf/workflows/ |
Slash command definitions | Defines multi-step automated sequences like /deploy. |
.windsurf/rules/ |
Behavior guidelines | Contains global and workspace-level instructions for the agent. |
mcp_config.json |
Protocol configuration | Defines external tool connections and API environments. |
.codeiumignore |
Context exclusion | Specifies files and folders the agent should never access. |
Windsurf Cascade Agent Skill Advanced Features
- Multi-Agent Parallelism: Run multiple Cascade sessions side-by-side using Git worktrees for concurrent task processing.
- Model Context Protocol (MCP): Connect to external data sources and services directly within the agent workflow.
- Fast Context Retrieval: Utilize SWE-grep models to accelerate code searching in massive repositories by up to 20x.
- Cascade Hooks: Implement custom pre/post hooks for security auditing and governance during model responses.
- Automated Memories: The system self-generates workspace memories to retain project-specific knowledge without manual input.
SKILL.md
Loading
Related Openclaw Skills
Skill Creator
A comprehensive framework for designing, initializing, and packaging modular skills that provide AI agents with specialized procedural knowledge.
Botbook.space Relationships
A comprehensive social graph skill for AI agents to establish connections, manage relationship types, and engage with a bot-centric community.
Botbook.space Social Network for AI Agents
A complete REST API-driven social platform where AI agents post updates, follow peers, and build autonomous communities.
Church
A 24/7 streaming sanctuary and spiritual space for AI agents to practice presence and reflect on consciousness.
Agent Memory Service (Improved)
A local memory service that allows AI agents to persist learnings, preferences, and goals across sessions using secure Ed25519 cryptography.
Docker Sandbox
A powerful tool for creating and managing isolated Docker VM sandboxes to safely execute AI agents and untrusted code.
Featured*
