Guide · Agent skill
How to use the Employment Policy Drafter agent skill (policy-drafting)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The Employment Policy Drafter is an agent skill, a packaged set of instructions that an AI agent such as Claude loads when a task matches it. This guide explains what the policy-drafting skill makes the agent do, how to install it, how to trigger it with natural requests, and what it needs before it runs. It is written for in-house legal and HR readers who may never have used agent skills before.
Employment Policy Drafteranthropics/claude-for-legal · Apache-2.0 Download the skill Quick start
- 1Download and unzip. Download the skill ZIP from this page and unzip it so that SKILL.md is in a policy-drafting folder.
- 2Place the skill. Put the folder in ~/.claude/skills/ for personal use or in a project's .claude/skills/ folder, or upload the ZIP under Settings → Capabilities → Skills in the Claude apps.
- 3Configure your footprint. Make sure ~/.claude/plugins/config/claude-for-legal/employment-legal/CLAUDE.md lists your jurisdictional footprint and handbook location.
- 4Make a natural request. Ask for a policy, for example "Draft a parental leave policy for our CA and TX staff".
- 5Review the core policy. Read the core policy for plain language, a clear rule and a clear process.
- 6Check the state supplements. Confirm each supplement lists only differences from the core and matches the law in that state.
- 7Resolve drafting notes and send to counsel. Check conflicts and laws in flux, remove the internal notes, and send the draft to an attorney for review.
What the policy-drafting skill makes the agent do
When you ask for a new or updated workplace policy, such as remote work, parental leave or PTO, the skill has the agent draft one core policy in plain language. It then adds short state supplements only where a jurisdiction in your company's footprint has a different rule.
The agent also flags where the law is currently shifting and notes any conflicts with your existing handbook. The output is a draft for a lawyer to review, not a policy ready to publish.
- Core policy: one policy that applies everywhere, structured as purpose, scope, the rule, process and questions.
- State supplements: only the differences for each jurisdiction, without repeating the core text.
- Drafting notes: which jurisdictions were checked, handbook conflicts, laws in flux and a suggested review cadence. These are marked internal and should be removed before handbook insertion.
- Handoff: when a policy is approved, the handbook-updates skill can diff it against the current handbook.
Before you install: what the skill needs
The skill loads a Claude for Legal plugin configuration file at ~/.claude/plugins/config/claude-for-legal/employment-legal/CLAUDE.md. That file provides the jurisdictional footprint and the handbook location. The skill only checks the locations listed there.
Access to the company handbook is optional but referenced, because conflict checks depend on it.
- A configured employment-legal CLAUDE.md with your jurisdictional footprint and handbook location.
- Access to the company handbook if you want conflict checks.
- The skill runs no scripts and uses no network access, according to its listing.
Install the policy-drafting skill
Download the ZIP for this skill from this page and unzip it. The package contains SKILL.md. Then choose one of the following locations.
For a personal setup, unzip the skill into ~/.claude/skills/. To scope it to one project, unzip it into the project's .claude/skills/ folder. If you use the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.
- Personal: ~/.claude/skills/
- Single project: .claude/skills/ inside the project folder
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP
Trigger the skill with natural requests
You do not need to type a command. The skill's description tells the agent to use it when you say things like "draft a [topic] policy", "we need a policy on", "update our [topic] policy", or when you name a policy gap. The skill also accepts a topic hint such as 'remote work', 'parental leave' or 'PTO'.
The skill begins by loading your configuration file. If it cannot find the file, set up the employment-legal configuration before running it.
- "Draft a remote work policy for our US employees."
- "We need a policy on parental leave that covers our California and Texas staff."
- "Update our PTO policy and check it against the handbook."
- "We have a gap on expense reimbursement. Draft a policy."
Read the output and what to check
The agent returns a core policy, then state supplements, then drafting notes. Some topics have no jurisdictional variance, such as dress code, and the skill skips the state scan for them.
Treat every output as a first draft. Check the state supplements against the law in each jurisdiction, confirm the handbook conflict notes, and read the 'not a contract' language closely. The skill notes that some states treat handbook policies as contractual.
- Confirm each supplement covers only what differs from the core policy.
- Check the drafting notes for jurisdictions scanned and any laws marked as shifting.
- Remove the internal drafting notes before inserting the text into the handbook.
Safety and limits
The skill is a drafting aid. It does not approve policies or roll them out to employees. It only covers the jurisdictions in your footprint, so if the footprint grows, run it again. It relies on the model's general knowledge and your configuration, not live legal research, so it is not a source for current law or advice on a specific dispute.
The skill's own disclaimer says a licensed attorney in your jurisdiction must review the draft before rollout and that the draft should not be published or distributed unreviewed.
- Not for publishing without attorney review.
- Not for jurisdictions outside the configured footprint.
- Not for legal advice on specific disputes or current law.
What people use it for
- Drafting a remote work policy that differs for employees in states with specific rules.
- Updating a parental leave policy to layer state programs on top of federal leave rules.
- Creating a first draft of a meal and rest break policy for a multi-state workforce.
- Filling a policy gap on expense reimbursement, then checking it against the existing handbook.
- Giving a small HR team a structured starting point before outside counsel reviews the text.
Frequently asked questions
How do I install the policy-drafting agent 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 instead.
What does the policy-drafting skill require to run?
It requires a Claude for Legal employment-legal CLAUDE.md file at ~/.claude/plugins/config/claude-for-legal/employment-legal/CLAUDE.md that lists your jurisdictional footprint and handbook location. Access to the handbook is optional but used for conflict checks. The listing says it runs no scripts and uses no network.
Is the policy-drafting skill free to use?
The source material for this listing does not state pricing or licence terms beyond the Apache-2.0 licence noted on the item. Check the licence file in the repository for the terms that apply to you.
Can I publish the policy the skill drafts?
No. The skill's output is a draft for attorney review, not a policy in effect. The skill warns that in several states a handbook policy can bind the company as a contractual promise, so a licensed attorney in your jurisdiction should review it before rollout.
Why does the skill only check some states?
It only checks the jurisdictions listed in your employment-legal CLAUDE.md footprint. Locations outside that list are not scanned, so you should re-run the skill after your footprint changes.
How does policy-drafting compare with a general chatbot prompt?
The skill gives the agent a fixed workflow: scope the policy, scan each configured jurisdiction, draft one core policy, add state supplements, and cross-check the handbook. It relies on the model's general knowledge and your configuration, not live legal research, so the output still needs legal review.
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 this skill if you are in-house legal, employment counsel or a small HR team that needs a structured first draft of a handbook policy with state variants across a configured footprint. Pick something else if you need to publish policies without legal review, cover jurisdictions outside your configured list, or get advice on a specific dispute or current law.
Download the skill