Plan Manager

Unified plan management system for coordinating OpenClaw Agent and Human plans. Use when managing both AI agent execution workflows and human task plans in a...

Security Scan

The skill is an instruction-only set of templates and a guide for creating plan files; its requirements and instructions are consistent with that purpose and it doesn't request credentials, external network endpoints, or install code.

Purpose & Capability

Name/description (plan management for agent and human tracks) matches the actual contents: templates and prose for creating files under plans/; there are no unrelated environment variables, binaries, or install actions.

Instruction Scope

SKILL.md only instructs creating/updating markdown files in a local plans/ directory and using provided templates. It does not mention reading unrelated system files, environment variables, or sending data to external endpoints. Note: the instructions assume the agent can write to the workspace filesystem.

Install Mechanism

No install spec and no code files to download or execute; lowest-risk instruction-only skill.

Credentials

The skill requests no environment variables, credentials, or config paths — proportionate to a templating/plan-management helper.

Persistence & Privilege

always is false and the skill doesn't attempt to modify other skills or system-wide settings. It will rely on the agent performing file writes when invoked, which is normal for a planning/template skill.

Assessment

This skill is internally coherent and low-risk: it provides markdown templates and tells the agent to create/update files under plans/. Before installing, consider: (1) the agent will need write access to your workspace to create these files—ensure you are comfortable with that; (2) review the templates for any accidental inclusion of sensitive data before committing to version control; (3) provenance is unknown (no homepage), so if you require an auditable source, request or host templates from a known repository. If you do not want the agent to autonomously modify files, restrict its permissions or review changes before applying.

Unified plan management system for coordinating OpenClaw Agent execution plans and Human task plans.

Overview

This skill provides a dual-track planning system:

Track	Purpose	Location
Agent Plans	Track OpenClaw Agent's execution workflows, tool sequences, and task progress	plans/agent/
Human Plans	Track human tasks, todos, and project milestones	plans/human/

Directory Structure

plans/
├── README.md                 # This file - overview and coordination guide
├── agent/
│   ├── plans-overview.md     # Agent's high-level execution board
│   ├── detailed/
│   │   ├── index.md          # Index of all detailed agent plans
│   │   └── XX-{plan-name}.md # Individual agent execution plans
│   └── archive/              # Completed agent plans
└── human/
    ├── plans-overview.md     # Human's high-level task board
    ├── detailed/
    │   ├── index.md          # Index of all detailed human plans
    │   └── XX-{plan-name}.md # Individual human task plans
    └── archive/              # Completed human plans

When to Use Which Track

Use Agent Track When:

• The OpenClaw Agent needs to execute multi-step tasks
• Tool sequences and execution workflows need documentation
• Tracking token usage and execution progress
• Breaking down complex user requests into actionable steps

Use Human Track When:

• The human user has personal tasks or todos
• Project milestones and deadlines need tracking
• Long-term goals and plans need organization
• Human decision points or approvals are required

Coordination Patterns

Pattern 1: Agent Assists Human Plan

Human creates a plan → Agent creates corresponding execution plan → Agent executes → Both update status

Example:

Human: "Plan a trip to Japan"
  ↓
human/plans-overview.md: "Plan Japan trip" (Status: In Progress)
  ↓
agent/plans-overview.md: "Research flights, hotels, itinerary" (Status: In Progress)
  ↓
[Agent executes research tasks]
  ↓
Both tracks updated to Done

Pattern 2: Agent Autonomous Execution

User requests complex task → Agent creates execution plan → Agent executes independently

Example:

User: "Refactor the auth module"
  ↓
agent/plans-overview.md: "Auth module refactor" (Status: In Progress)
  ↓
[Agent executes refactoring with detailed plan]
  ↓
Agent updates status to Done

Pattern 3: Human Oversight

Agent creates plan → Requests human approval → Human approves/modifies → Agent executes

Example:

User: "Design a new feature"
  ↓
agent/plans-overview.md: "Design new feature" (Status: Pending Approval)
  ↓
[Agent presents plan to human]
  ↓
human/plans-overview.md: "Review and approve feature design" (Status: In Progress)
  ↓
[Human reviews, approves]
  ↓
Agent proceeds to execution

Quick Start

For Agent Tasks:

1. Create plans/agent/plans-overview.md
2. Add execution items to the status table
3. Create plans/agent/detailed/XX-{plan-name}.md for complex tasks
4. Update plans/agent/detailed/index.md
5. Mark steps complete as execution progresses

For Human Tasks:

1. Create plans/human/plans-overview.md
2. Add task items to the status table
3. Create plans/human/detailed/XX-{plan-name}.md for complex tasks
4. Update plans/human/detailed/index.md
5. Update statuses weekly or as tasks complete

Templates

Templates are organized by track:

templates/
├── agent/           # Templates for Agent plans
│   ├── plans-overview.md
│   ├── detailed-plan.md
│   └── index.md
└── human/           # Templates for Human plans
    ├── plans-overview.md
    ├── detailed-plan.md
    └── index.md

Status Values

Status	Meaning
Pending	Plan created, not yet started
In Progress	Actively being worked on
Blocked	Waiting on dependency or approval
Done	Completed successfully
Cancelled	No longer needed

Tips

• Prefix plan IDs: Use A-1, A-2 for agent plans and H-1, H-2 for human plans when cross-referencing
• Link related plans: In the Notes column, reference related plans from the other track
• Archive completed plans: Move completed plans to the archive/ folder monthly
• Review weekly: Check both tracks weekly to ensure coordination
• Use consistent naming: kebab-case for all filenames (01-auth-refactor.md)

Example: Coordinated Web Development Project

Human Track (plans/human/plans-overview.md):

| H-1 ||***|| Launch new website ||***|| In Progress ||***|| High ||***|| Q2 2025 ||***|| June 30 ||***|| A-1, A-2 |

Agent Track (plans/agent/plans-overview.md):

| A-1 ||***|| Design database schema ||***|| Done ||***|| High ||***|| Week 1 ||***|| - ||***|| For H-1 |
| A-2 ||***|| Implement API endpoints ||***|| In Progress ||***|| High ||***|| Week 2-3 ||***|| - ||***|| For H-1 |

This structure ensures both human goals and agent execution are visible and coordinated.