Skip to main content
Glama

Server Details

Work Arbituro jobs from a coding agent: read the done-when, link the PR, claim what you checked.

Ownership verified
Status
Healthy
Uptime
15.4% over the last 24 hours
OAuth
Works in Glama
Last Tested
Transport
Streamable HTTP · MCP 2025-11-25
URL
Repository
writan-labs/arbituro-mcp
GitHub Stars
0
ChatGPT
–
Claude
–

TDQS

A3.9/5.0

Scored across 12 tools

Disambiguation4/5

Most tools target distinct actions: claim_done, file_result, link_pull_request, and propose_job_update are clearly separate verbs on the job lifecycle. The only soft overlap is among the read/history tools (job_history per-job changelog, what_happened project timeline, list_comments conversation), but each description pins its scope clearly enough.

Naming Consistency4/5

Nearly everything is consistent snake_case with a verb_noun shape (create_job, get_job, list_jobs, post_note, claim_done). A couple of outliers break the pattern — job_history is noun_noun and what_happened lacks a noun — but they remain readable and unambiguous.

Tool Count5/5

Twelve tools is well within the sweet spot and each maps to a concrete job/project operation (CRUD on jobs, comments, results, PR linking, search, history). No tool feels redundant or filler.

Completeness4/5

Core agent workflows are covered: create, read, list, comment, propose edits, file results, claim clauses, link PRs, and search. Notable gaps are assignment (get_job reports an owner but no tool assigns or claims a job) and direct status/close operations, though the propose/comment mechanisms let agents work around these.

Publisher details

Operator
Writan Labs · Publisher source
Vendor relationship
Not applicable
Trust center
Not available
Restrictions
Not available

Available Tools

12 tools
claim_doneClaim a done-when is metAInspect

Record that you believe one of a job's done-when clauses is now met, with your evidence. This does not tick the clause — a person decides that — it attaches your finding to the clause so they decide holding it. Say what you actually checked.

Inputs: job (required), clause (required), project (required), evidence (required).

create_jobFile a jobAInspect

File a new job in a project. It arrives unassigned and unagreed — nobody has said what done means for it yet, and nothing runs. Use it for work you found that nobody has written down.

Inputs: title (required), project (required), description.

file_resultFile a job's resultAInspect

File the document a job asks for, as Markdown, when get_job says its expected result is a document, a one-page memo or a slide deck. It becomes the job's result under the name of the person you act as, in that shape: there is no kind to choose. It is checked as it is filed, and refused with what to fix when a reader would see something wrong. Filing again replaces it with the next version, so send the whole document each time. This ticks no clause and does not finish the job: the job's judge is asked to finish it, and finishing is their verdict. A job whose result is a pull request, or that files nothing, is refused.

Inputs: job (required), project (required), markdown (required), title.

get_jobGet a jobA
Read-onlyIdempotent
Inspect

Get one job in full: its description, the plan a person approved for how to do it when there is one, the done-when clauses that define completion and the state each is in, who it is assigned to, the job it is part of, how many parts it has, and what has been spent on it.

Inputs: job (required), project (required).

job_historyRead a job's historyA
Read-onlyIdempotent
Inspect

What changed on one job and who changed it, most recent first. Use it to find out why a job looks the way it does before changing it.

Inputs: job (required), project (required), limit.

list_commentsRead a job's conversationA
Read-onlyIdempotent
Inspect

Read the conversation on one job: what people and agents have said about it, oldest first.

Inputs: job (required), project (required).

list_jobsList jobsA
Read-onlyIdempotent
Inspect

List this project's jobs in the order people set, one line each: key, title, type, status and its kind, owner, agent, start and due days, priority, parts finished, the job it is part of, how well it is written down (shaped n/3), the days since anything happened on it, and who it is waiting on and for what. Unfinished jobs unless status says otherwise, 50 at a time.

Inputs: project (required), after, limit, owner, +3 more.

post_notePost a noteAInspect

Leave a comment on a job, as yourself. It appears under your name in the job's conversation and notifies the people you @mention who can open the job. Naming an agent here starts nothing. Use it to report what you found, to ask a question, or to say what you are about to do.

Inputs: job (required), body (required), project (required).

propose_job_updatePropose a change to a jobAInspect

Propose a change to a job's title or description, or done-when clauses to add to it. Changes nothing: it raises a proposal on the job owner's queue, and they accept, edit or dismiss it. Use this rather than asking a person to make an edit for you — the proposal carries your wording, and one press applies it.

Inputs: job (required), project (required), title, doneWhen, +2 more.

what_happenedWhat happenedA
Read-onlyIdempotent
Inspect

The record of what has happened in a project recently — decisions, runs, approvals and changes, as one timeline. Use it to catch up before working, or to find out how something came to be the way it is.

Inputs: project (required), job.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      B
      maintenance
      Enables an AI agent to retrieve assigned tickets and their context, assign and mark tickets in progress, create pull requests that close tickets, and move tickets through review and done states while leaving Git actions to the agent and merges to humans.
      5
      MIT
    • A
      license
      Not graded
      quality
      B
      maintenance
      Gives AI coding agents a local, mechanical workflow authority: an agent's issue claim is recorded, every subsequent write and irreversible step (push, merge, tag, release) is re-verified against that record, and review verdicts are bound to exact commit bytes so a later push invalidates them. Refusals report what was read and what to do next, keeping an unreadable tracker distinct from a forbidden action.
      1
      MIT
    • A
      license
      A
      quality
      C
      maintenance
      Enables AI coding agents to submit work for external review before declaring a task done, receiving evidence-backed fix lists and verdicts based on git-measured changes and calibrated quality scoring.
      3
      MIT
    Try in Browser

    Glama MCP Gateway

    Add one secure layer between your agents and this server.