For engineers

Everything you used to ask the PM is already here.

The request, why it was decided, the acceptance criteria and what is still open. Claude Code and Codex read the same information. You never need to open the board.

Issue a token and register the MCP connection. Start with today's task.

A 30-second look at what changes for the people who build. Everything shown is the real product. No audio.
terminal
npm install -g pm-on-rails
pm-on-rails login
pm-on-rails mcp install

Connect Claude Code

login opens your browser for approval and stores the token in your local keychain. You need a workspace account to sign in. Restart Claude Code and tools like these are available.

Tools you will use most

start_task
Start a task. Returns the branch name and the line to paste into the pull request
read_scenario
Read an acceptance scenario (Given / When / Then)
get_traceability
Trace a task back to its request, and forward to what a change would affect
See every tool
(01)

The problem

Stuck before you write a line of code?

Finding the spec takes longer than writing the code

You cross-check meeting notes, Slack and issues before you can finally start.

The PM is busy, so you wait for an answer

You sent the question, and now you do not know what to do next.

The board still needs updating after the pull request

The code is done, but closing the task is still on your list.

(02)

How it works

Development that does not stop.

Leave what you do not know as a question and move on with what is decided. When the answer arrives, that part becomes a task you can pick up again.

  1. 01
    PM on Rails open questions. Pending questions, who they are for and whether they are answered. Work continues on decided tasks while you wait.

    Leave unknowns as questions

    Register a question the moment it comes up. You do not wait for the answer; you continue with the tasks that are already decided.

    You: Register the question

    About open questions
  2. 02
    PM on Rails Claude Code and Codex integration. The coding agent reads the request, acceptance criteria and task. What you used to ask the PM reaches the agent too.

    AI reads the same information

    Connected over MCP, Claude Code and Codex read the scenario and acceptance criteria before they implement. They do not mistake where to write or what done means.

    You: Implement

    About the MCP integration
  3. 03
    PM on Rails board. Tasks linked to pull requests move to done on merge. Progress updates without opening the board.

    Open a pull request and the task completes

    Link the branch and pull request, and passing CI is recorded as the scenario result. Merging completes the task. You stop opening the board.

    You: Open the pull request

    About the GitHub integration
(03)

The loop

From request to passing test, the same loop every time.

You are involved in the two highlighted steps. Requests and feedback are handled by the PM and the customer.

  1. 01RequestNotes and asks
  2. 02ScenarioAcceptance criteria
  3. 03BuildClaude Code / Codex
  4. 04TestCI results recorded
  5. 05FeedbackQuestions and decisions
Where you are involved
Where AI and the tool take over
(04)

From early users

What surprised the people who tried it.

It connects to CI and records the pass automatically. Because scenarios are in plain language, it is clear what counts as done.
An engineer at a development agency
I did not expect it to double as a direct line to the customer.
A lead engineer at a development agency

Summarized from interviews with early users.

Focus on the code

You get to focus on writing code.

The tool handles the PM work and you implement. Try it on your own project first.