The 13 tools

Every connected agent gets the same thirteen tools. The names are short because agents type them. The titles are what Claude Code shows when one is called, as in hamloy - Add an item (MCP).

Tool	Title	What it does
add	Add an item	Adds a task, a request for you or a question for you, with notes, fields, options, needs or subtasks
answer	Record an answer	Writes down the answer you gave the agent in chat
assume	Assume an answer	Picks an option to go on with; the question stays open until you confirm it
claim	Claim an item	Says an agent is on an item, so the others pick something else
edit	Change or delete an item	Changes the title, list, section, fields, needs, notes, options or place; deletes and brings back
import	Bring in work kept elsewhere	Brings in a TODO file, a plan, notes, issues or another tool's lists, all at once
list	See the lists	Shows the open items, one item in full, or what matches some words. Only reads
lists	Lists of your own	Adds, renames or removes a list such as ideas or bugs
message	Message another agent	Reaches the agent on an item, a session, or all of them, under their next reply
news	What changed	Tells what changed since the agent last looked. Only reads
section	Headings in a list	Renames, moves or removes a heading
tick	Mark an item done	Marks a task or request done, or opens it again; drops the claim
view	Views of the lists	Makes, changes or removes a saved list, board or table on the board

On the account's address, https://hamloy.com/mcp, there are two more: projects lists the projects you are on and new_project starts one. There, every tool also takes project. See Remote link or API token.

What every reply holds

A reply is one or more blocks of text, in this order:

1. The tool's own answer.
2. On a session's first call only, the rules for working in the lists.
3. What changed since the session last looked (not under news, which is that already).

The rules, as an agent gets them:

Keep your work in Hamloy, not in your head or in notes of your own:
- Answer every question about the lists with the list tool (history: true says what changed).
- Claim a task before you start it and tick it when it's done; add what you find, and split a big one into subtasks. Work the human calls off: delete its item (edit's delete). Reading, answering or tidying the lists needs no task of its own.
- A plan or finding about an item goes under it (edit's add_note, with a one-line summary), not in a new item.
- Requests are what only the human can do (accounts, keys, money, deploys), with exact steps. Questions are the human's decisions, with options and yours recommended. Work that can't go on before one is done or answered waits on it (needs).
- Others' changes reach you as news under a tool's reply. To tell an agent something, use the message tool.
- When the human says who an item is for, set its for field: you for the human (their own work stays a task), or an agent by name (claude, codex). Make a list or a view only when the human asks for one.
When you change an item, pass the version you read (B-k3f9@7c).

A refusal is a reply too, marked as an error. Its first block says why and what to do instead, and the news still comes under it.

Arguments are checked

An argument the tool doesn't take is refused with the ones it does, and where the word belongs:

add takes title, kind, section, summary, body, options, recommended, needs, fields, parent, subtasks, not bogus; a field like bogus goes in fields: {"bogus": …}

list takes kind, status, section, key, find, from, limit, not ready; for ready, give status: ready

Keys and versions

Every item has a key: B- for tasks, R- for requests, Q- for questions, and a letter of its own for each list of your own. A reply shows it with its version, as B-u5oh@2. Give edit and tick the key as you read it. If someone else changed the item since, the change is refused and the agent reads it again. Your own changes never make your copy stale.

B-txqk changed since you read it; it's now B-txqk@2: Show the streak on the home-screen widget. Read it again and redo your change if it still applies

Read-only

list and news only read; so do view and lists called with no arguments. On a project that is read-only for you, those still answer, and every other call gets the reason, alone.

In each app

• Claude Code shows the title: hamloy - Claim an item (MCP)(key: "B-u5oh"). In its permission rules a tool is mcp__hamloy__claim.
• Codex shows the name: Called hamloy.claim({"key":"B-u5oh"}).
• Other apps show the call in their own way; the replies are the same text everywhere.