The top of your workspace's dashboard has been refreshed and consolidated into four cards: test, test runs, latest pass rate, and flaky tests. Hover over any card to see a breakdown by test type, then click through to drill into any report with the relevant filters already applied.
- Pass rate and flaky tests, now on the dashboard: These two critical metrics are now front and center.
- Breakdown by type on hover: Hover over any card to see how tests, test runs, pass rate, or flaky tests break down across browser, mobile, API, and performance tests.
- One click to the full picture: Click a breakdown row to land in a more detailed report for that testing area.
How it works
Hover over a card (or tab to it and press Space or Enter) to reveal the per-type breakdown, like the Tests card above. Each row in the breakdown links to a pre-filtered report for that test type, so you can go from "flaky tests are up" to the ranked list of flaky tests in one click.
Get started with AI coding agents
Building tests or debugging failures with Claude Code, Codex, or another coding agent? A new panel on the dashboard (shown above) walks you through connecting your agent to mabl in as quickly as one step. You can dismiss it as needed, and restore it later from the same spot.
The panel links to mabl's skills repository on GitHub, which packages ready-to-use mabl skills for coding agents. Installing it takes one command for Claude Code and ChatGPT Codex, with other agent support documented in detail in the repo:
Claude Code
/plugin marketplace add mablhq/skills
/plugin install mabl@mabl
Codex
codex plugin marketplace add mablhq/skills
codex plugin add mabl@mabl
Once installed, your agent can query test results, investigate failures, and create or edit tests through the mabl cloud MCP server without leaving your terminal or IDE.