The TPR Framework is a cognitive closed-loop system designed to transform ambiguous requirements into verifiable, executable, and reviewable results.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install tpr-framework-main
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
Copy this prompt to OpenClaw to install it automatically.
Help me install tpr-framework-main using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
The TPR Framework (Think / Probe / Review) provides a unified methodology for navigating complex tasks. It functions as both a cognitive mindset and a formal execution process, leveraging a triple-department structure (Zhongshu, Menxia, and Shangshu) to ensure high-quality delivery. By centering all work on contractual benchmarks (GRV) and maintaining rigorous documentation, the framework ensures that no decision is made without systematic justification and user approval.
At its core, the framework emphasizes that orchestration should focus on scheduling rather than manual execution, and that documentation is the sole source of truth. This approach is essential for Openclaw Skills users who require professional-grade rigor in AI-driven project management and complex analytical workflows.
To integrate this methodology into your agent environment, declare your configuration in your AGENTS.md file:
# Example configuration for orchestrator agents
tpr_mode: full
can_spawn: true
model_config: config/tpr-model-config.md
Ensure all reference documents from the references/ directory are accessible to the agent for on-demand loading during specific workflow stages.
The skill organizes data through a strict documentation hierarchy to ensure transparency and traceability:
| Document | Phase | Purpose |
|---|---|---|
| DISCOVERY.md | Discovery | Requirement analysis and initial probing results. |
| GRV.md | Alignment | The 'Contract' containing Goals, Results, and Vital actions. |
| BATTLE-*.md | Review | Records of critical challenges and solution refinements. |
| output/* | Implementation | Final deliverables and generated technical assets. |
Loading
Automatically generate and edit publication-quality scientific vector diagrams from long-form text descriptions and reference images.

A high-performance utility to batch remove promotional metadata and ID3 tags from MP3 files.

A lightweight utility for AI agents to retrieve public IP addresses, geolocation data, and ISP information in real-time.

A professional planning tool for AI agents to structure signing workflows for business agreements like NDAs, MSAs, and SOWs.

An automated health monitor for AI coding agents that tracks work intensity and triggers natural rest reminders.

An AI-powered image generation skill that creates professional eCommerce posters and product detail pages via the Flyelep API.








































