Guide · Agent skill
How to use the REA Tool Design agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The rea-tool-design skill is a set of instructions that teaches an AI agent how to design or modify tools inside the REA project, an investigation toolkit that exposes commands through a CLI and an MCP (Model Context Protocol) interface. This guide explains what the skill makes the agent do, how to install it, how to trigger it, and what to check before you rely on its output. It is written for maintainers and contributors of the morluto/rea repository.
REA Tool Designmorluto/rea · MIT Download the skill Quick start
- 1Get the repository. Make sure you have access to the morluto/rea repository and its docs folder, including the tool-design, testing and CONTRIBUTING guides.
- 2Download the skill ZIP. Download the rea-tool-design ZIP from this page.
- 3Unzip into a skills folder. Unzip into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder to limit it to one project.
- 4Or upload in the Claude apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the ZIP.
- 5Ask for a design. Describe the tool, CLI or MCP contract change in plain language and ask for the boundary, input/result shape, Evidence semantics and nearest alternative.
- 6Review before implementing. Check the proposed design and verification plan, then ask for implementation limited to the scope you approve.
What does the rea-tool-design skill make the agent do?
The skill tells the agent to consult the project's tool-design guide for tool boundaries, contract semantics, provider ownership and discoverability. It also tells the agent to inspect the current contract and the nearest existing tool relevant to your question, because source code, documentation and live catalogs may describe different versions.
For a design request, the agent returns a proposed boundary, an input/result shape, Evidence semantics, a nearest alternative and a verification approach. It implements changes only when you ask, and it keeps to your stated scope and existing authorization.
- Design output: boundary, input/result shape, Evidence semantics, nearest alternative, verification approach
- Implementation: only when requested, limited to the scope you gave
- Implementation or verification: consults the testing guide for the affected consumer and provider lanes, and the contributing guide for generated-file ownership
What you need before installing
The skill requires access to the morluto/rea repository and its docs folder, which contains the tool-design, testing and CONTRIBUTING guides. Without those files, the agent has nothing to consult, and the skill will give little value. The skill is not a general tool-design guide and is aimed at people working on the REA codebase.
The skill's package contains a single file, SKILL.md. It does not run scripts and does not use the network.
- Access to the morluto/rea repository and its docs folder
- An agent environment that supports skills (Claude apps or Claude Code)
How to install the rea-tool-design skill
Download the ZIP for the skill from this page and unzip it. You have two options depending on how you use the agent.
For a personal install across all your projects, unzip the skill into ~/.claude/skills/. To limit it to one project, unzip it into that project's .claude/skills/ folder. In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.
- Personal (all projects): ~/.claude/skills/
- Single project: the project's .claude/skills/ folder
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP
How to trigger the skill with natural requests
You do not need a special command. Describe the design or change you want in plain language, and mention the REA tool, contract or provider involved. The skill is meant for design requests and for implementation or verification work, so state which one you want.
Asking for a design first is useful when you want a reviewed plan before any code is written. Ask for implementation only when you are ready for the agent to change files, and name the scope so it stays inside it.
- Design request: "Propose a design for a new REA investigation tool that returns provider capability results. Give the boundary, input and result shape, and Evidence semantics."
- Contract change: "Review how the existing CLI contract for this tool handles Evidence. Suggest the nearest alternative before changing anything."
- Implementation with scope: "Implement only the MCP contract change in the provider lane described in your design. Do not touch other tools."
- Verification: "Check which testing guide sections apply to this provider change and list the verification steps."
Safety notes: scripts and network
The skill does not run scripts and does not use the network, according to its description. The agent still edits files when you ask it to implement a change, so review diffs before committing. Because the skill tells the agent to use the existing authorization in your session, do not grant broader permissions than the task needs.
- Runs scripts: no
- Uses network: no
- Review all generated changes before merging
Tips for getting useful results
Name the specific tool or contract you are working on. The skill inspects the nearest existing tool, so a precise target gives a better comparison. Remember that the agent may find that docs, source and live catalogs disagree; treat that disagreement as something to resolve, not ignore.
Ask for the design and the verification plan together, so you can check the plan before any implementation starts.
- Name the exact CLI command, MCP contract or provider involved
- Ask for the nearest alternative to compare against your idea
- Request a verification plan alongside the design
- Keep implementation requests narrow and explicit
What people use it for
- Designing a new REA investigation tool and deciding where its boundary sits
- Reshaping a CLI or MCP contract without breaking existing consumers
- Adding or changing a provider capability and checking the Evidence semantics
- Getting a verification plan that points to the right testing guide lanes before coding
Frequently asked questions
How do I install the rea-tool-design skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. In the Claude apps, you can upload the ZIP under Settings → Capabilities → Skills.
What do I need to use the rea-tool-design skill?
You need access to the morluto/rea repository and its docs folder, which holds the tool-design, testing and CONTRIBUTING guides. The skill is built around those files, so it will not give much value without them.
Is the rea-tool-design skill free and what is its licence?
The item is listed under the MIT licence. Check the repository's licence file for the full terms.
Does the skill run scripts or use the network?
According to its description, the skill runs no scripts and uses no network. It consists of a single SKILL.md file. The agent can still edit files when you ask it to implement a change.
Can I use this skill outside the REA project?
It is not recommended. It relies on REA's own docs and file layout and is not a general tool-design guide, so it will give little value on other codebases.
What should I use instead if I just want to run REA investigations?
This skill covers designing and changing tools, not everyday use. If you only need to run REA for an investigation, use the project's own documentation rather than this skill.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use rea-tool-design if you maintain or contribute to the REA project and want a reviewed design, with a nearest alternative and verification plan, before changing CLI or MCP contracts or provider capabilities. Skip it if you are outside the REA codebase or only need to run REA investigations, since it depends on REA's own docs and does not cover everyday use.
Download the skill