Guide · Agent skill
How to use the Demand Letter Triage agent skill (demand-received)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The demand-received skill is a set of instructions that tells an AI agent how to do a first-pass triage of a demand letter: a formal claim that someone is owed money or must stop doing something. It is written for in-house legal teams, legal ops staff and business owners who want a structured read before deciding what to do. After this guide you will know what the skill does, how to install it, how to trigger it with a plain request, and what it needs set up first.
Demand Letter Triageanthropics/claude-for-legal · Apache-2.0 Download the skill Quick start
- 1Download and unzip. Download the skill ZIP from the directory page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder.
- 2Or upload in the Claude apps. In the Claude apps, open Settings, go to Capabilities, and upload the skill under Skills.
- 3Set up the config folder. Make sure ~/.claude/plugins/config/claude-for-legal/litigation-legal/ contains matters/_log.yaml and CLAUDE.md.
- 4Share the demand letter. Give the agent the path to the incoming letter or drop it into the session.
- 5Ask for a triage. Say something like "triage this demand" so the skill is triggered.
- 6Review the merit rating and options. Read the rating, the four response options and the recommendation, and check any flagged citations.
- 7Choose a handoff. Pick create matter, counter-demand, link to an existing matter, or standalone, and the agent acts on that choice.
What is an agent skill and what does this one do?
An agent skill is a folder containing a SKILL.md file. The file holds instructions that an agent such as Claude Code loads when a request matches the skill's description. This skill is named demand-received and sits in the litigation-legal part of the anthropics/claude-for-legal repository, licensed under Apache-2.0.
When you hand it an incoming demand letter, the agent reads the document and extracts the sender, recipient, delivery method, dates, demand type, specific asks, alleged facts, legal basis, threats and any settlement-communication framing. It then compares the letter against your matter log to spot repeat or related disputes.
It gives a merit rating of substantial merit, debatable, weak or frivolous and lays out four response options (substantive response, holding letter, settlement response, or ignore and preserve) with tradeoffs. It recommends one. It can also write a triage file and hand off to other legal intake skills. The skill is a structured first read, not a legal opinion.
What you need before installing
The skill assumes a specific local configuration. It reads a matter log and a practice file from a fixed folder, and the cross-check and house-tone steps will not work without them.
The folder it expects is ~/.claude/plugins/config/claude-for-legal/litigation-legal/ and inside it a matters/_log.yaml file and a CLAUDE.md file. CLAUDE.md is where risk calibration, your landscape notes and your demand-letter practice (house tone and response defaults) are read from.
You also need Claude Code or a compatible agent with file access. An optional legal research tool (Westlaw, CourtListener, Trellis, Descrybe or a firm platform) can be used to verify citations. Companion skills matter-intake, demand-intake and legal-hold are used for handoffs.
How do you install the demand-received skill?
Download the ZIP for the skill from the directory page and unzip it. You can install it for yourself across all projects or for a single project.
For a personal install, unzip the folder into ~/.claude/skills/. For a project-only install, place it in the project's .claude/skills/ folder.
If you use the Claude apps instead, open Settings, go to Capabilities, and upload the skill under Skills.
- Personal: ~/.claude/skills/
- Single project: .claude/skills/ inside the project
- Claude apps: Settings → Capabilities → Skills, then upload
How do you trigger the skill with a natural request?
You do not need to memorise a command. The skill is triggered by requests like the ones its description lists. Share the incoming letter and say what you want.
Good triggers include "we got a demand letter", "triage this demand", or sharing an incoming demand and asking for an evaluation. You can also point at a file path, since the skill's argument hint accepts a path to the incoming document and an optional custom slug.
- "We got a demand letter from a supplier. Triage this demand: /path/to/letter.pdf"
- "Triage this demand and tell me whether it has merit."
- "Here is a demand we received this morning, evaluate it."
/demand-received [path-to-incoming] [--slug=custom-slug]What happens during a triage run?
The skill works through a fixed sequence. It reads the incoming document, loads the matter log and CLAUDE.md, then extracts the fields listed above. The cross-check looks for a direct match (same counterparty), a type match (similar past matter with this counterparty, including closed ones), or a subject overlap such as the same contract or project.
Merit is assessed on facts, legal basis, each side's likely strength, whether the damages ask is proportionate, and the sender's leverage. Any cited law is flagged for your verification rather than checked by the skill itself.
Deadlines are triaged separately: their stated deadline, your internal decision deadline, and any legal deadlines such as limitation periods or contractual cure periods.
- Read the letter and extract fields
- Cross-check _log.yaml for related matters
- Load CLAUDE.md for calibration and house practice
- Rate merit and present options with a recommendation
- Write triage.md and copy the incoming letter into an inbound folder, then hand off if you choose
Safety, scripts and network use
This skill does not run scripts and does not use the network. It works on files you give it and on the configuration folder described above.
It writes its output under ~/.claude/plugins/config/claude-for-legal/litigation-legal/inbound/[slug]/, creating triage.md and a copy or link of the incoming document there.
Citations are tagged with their source, such as Westlaw, CourtListener or a web search marked for verification. If a research query returns thin results, the skill stops and asks you how to proceed rather than filling the gap silently. Verify any conclusion with a licensed lawyer before relying on it.
- No scripts are run by this skill
- No network access is used by this skill
- Output is written to the inbound folder under your config path
What people use it for
- Triage a payment demand from a supplier and decide whether to respond substantively or send a holding letter.
- Check whether a new demand comes from a counterparty who already has an active or closed matter in your log.
- Prepare a settlement-response approach for an early-resolution demand, with notes on settlement-communication protection.
- Produce a consistent first-pass intake record for each demand so legal ops can compare them.
- Get a structured read of a breach or cease-and-desist letter before speaking to outside counsel.
Frequently asked questions
Is the demand-received skill free to use?
The source does not state pricing or licence terms beyond the Apache-2.0 licence for the repository. Check the licence and the directory page for current terms before use.
What do I need to install demand-received?
You need Claude Code or a compatible agent with file access, the skill folder unzipped into ~/.claude/skills/ or .claude/skills/, and a config folder at ~/.claude/plugins/config/claude-for-legal/litigation-legal/ containing matters/_log.yaml and CLAUDE.md.
Does the skill give legal advice?
No. It is described as a structured first read, not a legal opinion. It flags cited law for verification, so a licensed lawyer must check any conclusions before you rely on them.
Why won't the cross-check or house-tone steps work?
The skill assumes the matter log and CLAUDE.md practice file exist at the expected path. Without them the portfolio cross-check and house-tone steps cannot run as designed.
Does the skill send anything or access the internet?
The skill does not run scripts and does not use the network. It reads your files and writes triage output under your config folder.
How does it compare with other ways of triaging demand letters?
The skill's distinguishing features are the portfolio cross-check against your matter log, the four structured response options, and handoffs to matter-intake and demand-intake. A manual process or a general chat prompt will not read your matter log or write triage files in this format unless you set that up yourself.
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 an in-house legal team, legal ops group or business owner who receives demand letters and wants a consistent first read with a portfolio cross-check. Pick something else if you need final legal advice, do not have the matter log and CLAUDE.md configuration in place, or cannot have a lawyer verify the cited authorities.
Download the skill