How it works
- Connect the Flagsmith MCP server at https://mcp.flagsmith.com to your AI assistant or coding agent and sign in with OAuth, or use an API key header where your client runs without a browser, noting that experimentation in Flagsmith is currently in Beta.
- Install the ExpLint MCP at https://explint.com/mcp to enable your agent to audit tests.
- Run the prompt below to audit these plans before the work starts.
How the check runs
- It reads the experiment's name and hypothesis, the multivariate flag behind it with the current value as control, the rollout share entering the experiment beside the variation split, and the primary metric with its expected direction.
- It sends that text to ExpLint with
explint__lint, which returns errors, warnings, and passes, each finding naming the rule behind it. - You edit the experiment in Flagsmith and ask for a fresh read, keeping the findings in chat because the MCP tools change setup fields without storing a verdict on the experiment.
Copy and paste this prompt
Change the words in square brackets to name your own project, board, or test.
Prompt for your assistant
Use the Flagsmith MCP tools to read the experiment [experiment name]. Build one text from its name and hypothesis, variants with the control marked, rollout share and variation split, and primary metric with its expected direction. Lint it with explint.com against project [your ExpLint project name]. Show each error and warning with its Flagsmith field. Do not launch or edit the experiment yet. Suggest fixes here and re-check after I update Flagsmith.
What ExpLint looks for
ExpLint ships over 24 rules, and a project switches each one on or off or rewrites it in plain English. These are the rules the plans this guide reads break most often. 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 | An experiment that only claims the change will help gives the setup nothing a finished result could contradict. |
| variants/control-is-described | error | The current flag value is the control, and the plan should say what that value shows people today. |
| randomization/states-the-split | error | Rollout share and variation split are separate numbers in Flagsmith, and the plan should show both so the reader knows how many people enter and how they divide. |
| decision/has-criteria-before-launch | error | The setup names no decision field, so the plan should record what result would roll the flag out and what result would turn it off. |
