Lint Trello cards in your backlog with explint.com
Read experiment cards from a Trello board, check each card with explint.com in the same assistant chat, and write the findings back onto the card so the next person who picks it up can see what needs fixing.
What you need before you start
- A Trello board where experiment ideas live as cards, with a list for ideas that are still being shaped and a list for cards that are ready.
- An assistant that can hold two MCP connections at once, such as Claude, ChatGPT, Cursor, or Claude Code.
- An A/BLint account with a project your team can tune, starting from the default project.
- Cards that carry their plan in the description, because a card that only holds a title and a link gives the checker almost nothing to read.
Connect Trello and A/BLint
- Add Trello's official MCP server in your assistant at https://mcp.trello.com/v1 and sign in, then approve the read and write access it asks for board by board if it offers you that choice.
- Add A/BLint in the same assistant at https://explint.com/mcp and finish the browser sign-in, or use an API key from the MCP page in your A/BLint workspace if your assistant asks for a key instead.
- Ask the assistant to name the boards it can see and the lists on the board you use for experiments, because that confirms the connection is reading the board you meant.
- Say the name of your Ideas list and your A/BLint project in the chat, so the assistant knows which cards to read and which rules to use.
How the check runs
- The assistant reads the cards in your Ideas list, opening each card for its description, checklists, and labels.
- It joins the card title, description, and checklist items into one piece of text and sends that text to A/BLint with ablint__lint.
- It brings back a short report for the whole board that groups cards by what they are missing, which makes patterns easy to spot when the same gap shows up on card after card.
- You pick the cards worth saving, and the assistant drafts stronger descriptions in the chat for your approval.
- When you approve a draft, it writes the findings and the draft back to the card as a comment, and it can move the card to your Ready list once a re-check comes back clean.
Copy and paste this prompt
Change the words in square brackets to name your own project, board, or test.
Use the Trello MCP tools to read every card in the [Ideas] list on the board [board name]. For each card, lint the title, description, and checklist with explint.com against the project [your A/BLint project name]. Show me each card name with its errors and warnings, grouped by what is missing most often across the board. Do not change any card yet. When I approve a card, add a comment with the findings and a suggested rewrite, then lint the rewrite and tell me if the errors are cleared.
What A/BLint looks for
A/BLint scores the text against the rules switched on in your project. The default project uses the rules below, and your team can turn rules off, change how strict they are, or add its own rules in plain English. An error means the test cannot answer its question as written, and a warning means the result will be hard to judge when it comes in.
| Rule | Level | In plain words |
|---|---|---|
| hypothesis/is-falsifiable | error | The idea makes a clear guess that a result could prove wrong, instead of saying the change will improve things in general. |
| hypothesis/names-metric-and-direction | error | The guess names the number it expects to move and which way it should move, such as a rise in checkout completion. |
| hypothesis/is-based-on-evidence | error | The write-up points to something real behind the idea, such as research, usage data, support tickets, or an earlier test. |
| metrics/has-one-primary-metric | error | One main number decides the test, so the team agrees before the test starts which result matters most. |
| guardrails/are-listed | error | The write-up names the numbers that must not get worse, such as speed, errors, refunds, or people leaving. |
| duration/covers-full-business-cycles | error | The test runs long enough to cover full weeks, so a busy Monday and a quiet Sunday both count in the result. |
An example you can compare against
Before
Homepage banner test
Test a new banner on the homepage. See the design in the attached file. Check if it performs better after launch.
After, checked again
Test a free delivery banner on the homepage
- Guess
- If we add a banner that says orders over 40 dollars ship free, then the share of visits that end in an order will rise, because exit surveys keep naming delivery cost as the reason people leave without buying.
- Evidence
- In last month's exit survey, 34 percent of people who left named delivery cost, and orders just under 40 dollars are our largest group of abandoned baskets.
- Main number
- Share of visits that end in a paid order.
- Guardrails
- Average order value must not fall by more than 3 percent, and refunds must not rise.
- Split
- 50/50 for two full weeks, including two weekends.
Still flagged on the rewrite
- errorguardrails/have-launch-thresholdsThis guardrail has no limit written as a number the team can check.
- warningpower/has-a-sample-size-calculationThe run has a split and a length, and no count of visits or size working behind it.
The stronger card says what the banner promises, which number will judge it, and why the team believes it, all in the card itself. Keeping the guess and the numbers on the card matters on Trello because cards move between lists and people, and the reasoning needs to travel with the work.
Fix what it finds and check again
Start with the errors, and have the assistant suggest each rewrite in the chat before anything is written back to Trello, so wording your team already agreed stays yours to approve. Rewrite the card so a teammate reading it cold could say what will change, what number will judge it, and what would stop a launch, then lint the same card again and ask which rules now pass that failed before.
Limits and things to check yourself
- Trello's MCP server lets assistants archive cards and lists, and permanent deletes are not supported through it, so a mistaken write is usually easy to undo from the card activity log.
- Pictures, attachments, and linked sheets carry detail A/BLint cannot read, so paste the words that matter into the card description before you lint.
- Boards with hundreds of cards in one list are better checked in smaller groups, such as one month of ideas at a time, so the report stays readable.
- A passing score says the write-up covers what the rules ask for. It cannot say the idea is worth testing, that the traffic exists to reach a clear result, or that the change is safe to build, and those calls stay with your team.