A structured multi-phase strategic planning skill for AI product managers and coding agents to map evolutionary milestones from MVP to high intelligence.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install workflow-dlc-agent-phasing
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 workflow-dlc-agent-phasing 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 Agent Phasing?
The agent-phasing skill implements a reliable, gate-based framework for planning AI agent products across four critical progression phases (P0 to P3). Rooted in the One Click methodology, it balances aggressive MVP scoping with stringent architectural baselines. The core philosophy dictates that earlier phases should never sacrifice foundational infrastructure for speed, thereby protecting the project from crippling technical and architectural debt.
By deploying this skill within Openclaw Skills environments, developers and product owners can systematically map out feature sets while declaring explicit boundaries for what will not be built in each iteration. It ensures every deployment phase maintains verifiable data collection, protocol alignment, and governance standard operating procedures, shifting the development cycle away from guesswork toward structural predictability.
Agent Phasing Use Cases
The skill is triggered under the following scenarios:
- When a product manager or developer issues commands like "Agent planning", "P0 P1 phasing", or "Agent roadmap".
- During the initiation of automated
workflow-startrouting tasks requiring project roadmap structure. - When scoping out complex AI multi-agent workflows that risk creeping in scope or introducing technical debt.
How Agent Phasing Works
The operational workflow adheres to a rigid sequential logic:
- Route and Parse: The skill triggers via natural language queries or protocol hooks and reviews the overall project ambitions against structural lesson assets.
- P0 Foundation Execution: Establishes an end-to-end single-scenario framework with mandatory logging, rule-engines, and complete PageContext protocol definitions.
- P1 Closed-Loop Enforcement: Introduces flat knowledge repositories and analytical summaries to ensure the feedback loop operates seamlessly.
- P2 Expansion Verification: Validates legal and regulatory structures while extending features safely across multi-scenario domains.
- P3 Intelligence Layering: Introduces advanced autonomous action algorithms, real-time sync systems, and granular tracking playback mechanisms once structural baselines are stable.
Agent Phasing Setup
To integrate this workflow into your workspace, implement the skill configuration using your local setup:
# Install the agent skill within your directory profile
openclaw skill install agent-phasing
# Run a test initialization simulation for an initial P0 architecture blueprint
openclaw skill run agent-phasing --trigger "Agent roadmap"
Agent Phasing Data Schema & Taxonomy
The planning sequence evaluates scope blocks and writes progress records out to structured JSON logs according to this model:
| Property | Type | Description |
|---|---|---|
timestamp |
String | ISO 8601 formatted date and execution timeframe |
skill |
String | Identifies the execution origin block (agent-phasing) |
current_phase |
String | The target phase boundaries evaluated (P0, P1, P2, or P3) |
mvp_scope |
Object | Map container containing precise arrays for must_do and not_doing lists |
architecture_baseline |
Object | Explicit configuration flags verifying protocol, logging, and rules engine status |
phase_completion_criteria |
String | Quantifiable outcome description necessary to pass the phase gateway |
Agent Phasing Advanced Features
- Gate-Based Hard Stops: Automatically rejects scope suggestions that defer critical system logs or bypass legal-compliance alignment.
- Architecture Framework Enforcement: Mandates uniform API and context schema layout rules starting directly from the P0 baseline map.
- Scoping YAGNI Pruning Engine: Automatically flags feature creep patterns, stripping non-essential elements out of current sprint targets into future phase arrays.
- Structured Output Delivery: Directly outputs execution documents, MVP checklists, structural layout dependencies, and JSON state files.
SKILL.md
Loading
Related Openclaw Skills
Agent Scenario Positioning Skill
An expert AI agent workflow designed to precisely position an AI product's interaction model, core user scenarios, and data contexts, eliminating low-value chat behavior.
Cost Guard
Cost Guard is an AI-driven guardrail skill designed to optimize prompts, estimate costs, and intercept high-consumption workflows before they drain your token budget.
Companion AI Persona
An interactive agent skill that enables natural language switching between diverse boyfriend and girlfriend personality archetypes for simulated companion roleplay.
AI Product Manager Assistant
A robust AI co-pilot that automates PRD drafting, competitive analysis, data-driven diagnostics, and agile sprint alignment for product managers.
Agent Learning
A closed-loop execution framework that turns raw operational logs into structured agent experiences through automated AI analysis and human-in-the-loop review.
Agent Interaction Workflow
A design-oriented developer skill for compressing 6-step conversational dialogues into high-performance 3-step structured interactions for AI agents.
Featured*
