How to use the Infringement Triage agent skill

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

An agent skill is a folder containing a SKILL.md file with instructions that an agent such as Claude Code loads when a request matches the skill. This one is named infringement-triage and is invoked as /ip-legal:infringement-triage within the ip-legal plugin.

When you give it the facts of a possible infringement, the skill asks which right is at issue, runs common intake questions about party posture, jurisdiction, timing and exhibits, and walks the factors for that right. It then produces a flag list showing what cuts toward the senior party, what cuts toward the accused and what is mixed. It writes the triage memo to a matter folder or practice outputs folder and ends with recommended next steps.

• Trademark: confusion test, dilution if the mark is famous, and false advertising if a comparative claim is made.
• Copyright: ownership, registration, access, substantial similarity, fair use and DMCA safe harbor if applicable.
• Patent: a claim-chart first pass, literal infringement and doctrine of equivalents, indirect and divided infringement, and invalidity defenses to consider.
• Trade secret: secrecy, reasonable measures, misappropriation, and preemption and reverse-engineering flags.

How do you install the infringement-triage skill?

Download the ZIP from the skill's page on this site and unzip it. The source package for this skill contains a single file, SKILL.md. Place the unzipped skill folder in one of these locations:

You can install it for yourself in ~/.claude/skills/, or for a single project in that project's .claude/skills/ folder. Alternatively, in the Claude apps you can upload it under Settings → Capabilities → Skills.

• Personal install: ~/.claude/skills/ (available across your projects).
• Project install: .claude/skills/ inside the project folder.
• Claude apps: Settings → Capabilities → Skills, then upload the skill.

What do you need before it works?

The skill is designed to run with Claude Code or a compatible agent that has the claude-for-legal ip-legal plugin installed. It reads a practice profile at ~/.claude/plugins/config/claude-for-legal/ip-legal/CLAUDE.md. If that file still contains [PLACEHOLDER], the skill stops and points you to /ip-legal:cold-start-interview.

If you have not configured the profile, you can answer "provisional" to run the skill against generic defaults: US jurisdiction, middle risk appetite, lawyer role and no playbook. Provisional output is tagged [PROVISIONAL] so you can tell it apart from tailored output.

Optional integrations such as CourtListener or Solve Intelligence can be referenced in the practice profile. They let the triage cite case law, prior rulings or prior art. The skill does not require them.

How do you trigger it with a natural request?

Invoke the skill with the slash command and describe the facts. Include the right at issue if you know it. If you leave the facts out, the skill asks which right is at issue and for the facts.

If more than one right is in play, the skill runs a separate triage for each rather than blending them. A single set of facts can implicate trademark, copyright and trade secret at once.

Ask in plain language about a knockoff, a copycat product or a possible leak of your technology. Phrases such as "is this infringing our trademark" or "should we pursue this copycat" describe the situations the skill is written for.

Worked examples

The source gives two example requests. The first is a trademark question about a competitor product name. The second is a trade secret question about a former engineer. A third form runs the command with no arguments so the skill asks you for the right and the facts.

The skill will not tell you the names are confusingly similar or that the engineer misappropriated anything. It will list the factors and their direction and suggest next steps for the attorney.

What safety limits should you know about?

The skill is built never to conclude. Acting on a triage without attorney review can lead to fee awards, Rule 11 sanctions, declaratory-judgment actions and, for patents, treble damages. Treat the output as a structured first pass for a lawyer, not a decision.

According to the package metadata, this skill runs no scripts and uses no network access. Its instructions are contained in SKILL.md. Check the configuration and any integrations you connect before you rely on citations they produce.

Output can include a suggested offer to draft a cease-and-desist or takedown, but the skill does not draft these automatically.

• Do not send a letter, refuse to stop, file suit or decline to act based on the triage alone.
• Expect it to be US-focused unless your practice profile is configured for other jurisdictions.
• Under-calling a conflict and over-calling it carry different risks, so the skill is tuned to flag rather than conclude.

Tips for better triage output

Give concrete facts: the marks or works involved, the goods or classes, when you learned of the activity, and who the parties are. Say which side you are on, because the skill asks whether you are the senior party or the accused.

Configure the practice profile first. Your enforcement posture, registration jurisdictions and approver names change how the output is framed and where it routes.

• Name the right if you know it; otherwise let the skill ask.
• Open a matter workspace if the triage may lead to a cease-and-desist, so the documents stay together.
• Read the flag list for direction, then take the open questions to your attorney.