How to use the Employment Profile Customizer agent skill

What is an agent skill, and what does this one do?

An agent skill is a folder containing a SKILL.md file that gives an AI agent a specific set of instructions it can follow when a request matches. This one, customize, is a guided editor for an employment practice profile that the cold-start interview created earlier.

When you ask to change one item, the agent reads your current config, shows you the present value, asks for the new one, explains which other employment tools will be affected, confirms the change and writes it to the config file. It also flags internal contradictions, such as a California footprint combined with an aggressive non-compete posture, and warns when a change weakens safeguards like the citation check or the [review] flag.

• Customizable areas: company details, jurisdictional footprint, risk posture, people and escalation contacts, hiring review, termination review, handbook, investigation preferences, workflow and integrations.
• Changes to shared company details (name, industry, jurisdictions, practice setting, stage) are written to company-profile.md and affect all 12 plugins.
• It never deletes a section. Removing a jurisdiction means marking it as not currently staffed so its rules can be restored later.

What do you need before installing?

The skill only works on an existing profile. Before you install it, confirm that you have run the cold-start interview from the claude-for-legal employment-legal plugin, which is /employment-legal:cold-start-interview. If setup has not been completed, the skill stops and asks you to run it first.

The agent also needs read and write access to files under ~/.claude/plugins/config/claude-for-legal/. The skill does not run scripts and does not use the network.

• An existing claude-for-legal employment-legal plugin profile created by the cold-start interview.
• Read and write access to ~/.claude/plugins/config/claude-for-legal/ files.
• The package includes one file: SKILL.md.

How do you install the customize skill?

Download the ZIP from the directory page and unzip it. You can install the skill for yourself or for one project.

For a personal install, unzip it into ~/.claude/skills/. For a project install, unzip it into the .claude/skills/ folder inside your project. In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.

• Personal: place the skill folder in ~/.claude/skills/.
• Project: place the skill folder in .claude/skills/ within the project.
• Claude apps: upload the ZIP under Settings → Capabilities → Skills.

How do you trigger the skill with a natural request?

You can invoke the skill with the command /employment-legal:customize, or with a plain-language request. The skill's description lists phrases such as "change my [thing]", "add a jurisdiction", "update my profile", "edit my config" and "customize". You can also add a section name or a description after the command, since the argument hint accepts a section name or a description of what you want to change.

Describe one change at a time. The skill is built to adjust one item without re-running the whole interview.

• "Add Washington to my jurisdictions."
• "Change my risk posture to conservative."
• "Update my severance framework to 4 weeks per year."
• "Edit my config so the escalation contact is our people team lead."

What happens during a change?

The agent first reads CLAUDE.md in the employment-legal config folder, and company-profile.md one level up. It then lists the profile by section with a one-line summary of each current value. After you name what to change, it shows the current value, asks for the new one, and explains downstream effects before writing anything.

For example, adding Washington to the footprint means /wage-hour-qa and /termination-review start applying WA rules, /handbook-updates prompts for a WA supplement, and /hiring-review flags non-compete attempts in WA as unenforceable. Changing risk posture from middle to conservative means more terminations are flagged for escalation and restrictive covenants are reviewed more strictly.

• Expect the agent to ask for confirmation before saving.
• If a change conflicts with other settings, the agent will point out the tension.
• Changing a [review] flag or citation check requires the trade-off to be explained first.

What safety and accuracy limits should you know?

The skill edits configuration only. It does not verify the law and does not replace a licensed attorney. The pre-flight citation check, source attribution tags and [verify] tags on cited statutes are meant to stay in place, and the skill warns before any change that weakens them.

Because the skill has no scripts and no network use, its actions are limited to reading and writing the config files listed above. Review the profile after any change that affects jurisdictions or risk posture.

• Edits configuration, not legal advice.
• No scripts run and no network access is used.
• Keep the citation check, attribution tags and [verify] tags in place.

What tips help you use it well?

Keep changes small and specific, and read the downstream explanation before confirming. If you are unsure which section holds a setting, ask the skill to show the customizable map first. Run the command again whenever a single item needs updating.

• Start by asking for the map to see all sections.
• Name the section you want, such as "People" or "Termination review".
• Check the downstream effects list before you confirm.