The Plan Skill empowers AI agents to generate structured, step-by-step implementation blueprints and interactive sub-agent plan-deepeners before writing code.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install holbot-plan
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 holbot-plan 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 Plan Skill?
The Plan Skill is a core architectural framework designed for engineering-grade task orchestration within the ecosystem of Openclaw Skills. Grounded in the principle of 'think first, build later,' this skill defines the precise implementation structure (the how) before any actual code modification occurs. It helps developers and AI agents avoid short-sighted implementations by enforcing a rigorous, context-aware planning phase. Let your AI agent run research, map repository files, and model workflows systematically before diving into active execution.
By leveraging the Plan Skill, users can ensure their Openclaw Skills projects are organized, context-aware, and thoroughly reviewed. It decomposes highly complex goals into isolated, sequential, and testable steps, bridging the gap between high-level prompt requirements and executable code updates.
Plan Skill Use Cases
- Decomposing Multi-Step Tasks: Break down complex features, automation workflows, research topics, or architectural changes into orderly execution phases.
- Repository-Aware Implementation Mapping: Automatically scan existing codebases, workspace structures, and prior learning modules to craft context-rich development blueprints.
- Interactive Plan Deepening: Use sub-agents to detail execution strategies, discover edge cases, and add highly technical implementation requirements to existing high-level plans.
- Scope Definition and Boundary Setting: Explicitly list project inclusions, exclusions, and test scenarios to minimize scope creep in automated workflows.
How Plan Skill Works
- Scope and Context Initialization: The skill triggers on commands like 'plan this' or 'create a plan'. If the input is ambiguous, it requests brief clarification before bootstrapping.
- Repository and Workspace Research: The agent inspects local codebase structures, checks previous design choices via workspace memory systems, and conducts external web searches if third-party dependencies exist.
- Structured Plan Generation: A detailed planning document is structured containing core sections: Context & Problem, Goals, Approach, Scope, File Maps (using repo-relative paths), Implementation Steps, Acceptance Criteria, Risks, and Test Scenarios.
- Deepening and Verification: If 'deepen the plan' is requested, the system launches interactive sub-agents to analyze and enrich individual sections with edge cases and rigorous execution paths.
- Export and Serialization: The generated plan is presented for review and can be written as a clean Markdown file under
/home/HolBot/plans/for version-controlled tracking.
Plan Skill Setup
To enable and execute this skill within your ecosystem of Openclaw Skills, invoke it directly within your agent terminal or integrate it into your automated agent profiles.
CLI Activation
Simply call the skill using the registered slash command:
/plan [optional: task description, requirements, or file path to deepen]
Configuration
Ensure your agent environment is configured to write to the designated plan storage directory. You can pre-configure the destination directory in your environment variables or global skill configuration:
export OPENCLAW_PLANS_DIR="/home/HolBot/plans"
Plan Skill Data Schema & Taxonomy
The Plan Skill structures and saves plans using standard Markdown format. It records precise metadata and execution hierarchies to ensure compatibility with other Openclaw Skills.
Export Directory Structure
Plans are archived in the following format:
| Directory | Filename Pattern | Description |
|---|---|---|
/home/HolBot/plans/ |
YYYY-MM-DD-[slug].md |
Version-controlled plan archiving |
Plan Document Schema
Every plan file uses a highly structured Markdown schema to organize its sections:
- Context & Problem: Outlines current limitations and high-level project targets.
- File Map: Lists all affected repo-relative paths.
- Implementation Steps: Numbered steps including dependencies and sequence estimates.
- Acceptance Criteria & Test Scenarios: Verifiable tests mapped directly to feature units.
Plan Skill Advanced Features
- Interactive Plan Deepening: Spawns parallel sub-agents to interrogate, expand, and detail specific sections of a drafted plan.
- Memory Integration: Inspects
MEMORY.mdto load learnings, past mistakes, and architectural constraints automatically. - Repo-Relative Path Enforcement: Guarantees zero absolute paths in generated files, ensuring perfect portability across machines.
- Sub-Agent Orchestration: Multi-agent handoff workflows that review planning documents before handing them to implementation agents.
- Seamless Openclaw Skills Compatibility: Integrates directly with code-generation and test-running plugins to automate execution of completed plans.
SKILL.md
Loading
Related Openclaw Skills
Lingxi AI Note Recording Query
The Lingxi AI Note Recording Query skill enables AI agents to query, retrieve, and analyze sales and meeting voice recordings and their detailed transcription insights.
1688 Supply Chain API Procurement Skill
Automate product sourcing, inquiries, and pricing analysis on 1688 with secure raw JSON pipelines that preserve LLM context.
agent-plus
A comprehensive configuration framework to define structured AI agent identities, behavior-driven voices, dynamic adaptation rules, and robust operation boundaries.
Testing Plus
Testing Plus is a versatile software verification suite that equips projects with multi-language testing configurations, automated CI/CD workflows, mutation testing, performance load testing, and vulnerability diagnostics.
Room Status Query
A developer tool that queries real-time and cached meeting room availability and recently released slots.
Boss Fight Planner Skill
A structured game design utility providing repeatable templates, checklists, and safe defaults for planning epic boss encounters.
Featured*
