Run QA from your coding agent

Run checks from your terminal, coding agent or CI. Get results and evidence your agent can use to fix bugs.

Work from your agent

Run tests and read findings without leaving your coding workflow.

Check pull requests

Verify the exact deployed change from CI.

Get evidence for fixes

Read results, recordings and reproduction steps from the same runs your team sees.

1

Connect your agent

Install the Composal skill so your coding agent can use the CLI.

2

Run your checks

Run Scenarios from your terminal or verify a pull request from CI.

3

Review and fix

Read the results and evidence. Give your coding agent the details it needs to fix what broke.

What you can do with it

  • Stop CI when a check fails

    The command fails unless every Scenario passes, so your job stops on a failed or blocked run.

  • Let your agent take the next step

    Each failure says what to do next. Your agent does it, or tells you when it needs you, like signing in again.

  • Download the evidence

    Save a recording, trace or screenshot to a file. It never overwrites a file you already have.

  • Keep only what you need

    Trim results before they reach your agent, and see whether anything was cut.