Find what to test
For AI agents: this page is Markdown at https://squiggle.sh/docs/agentic-flows/coverage.md. Every docs page is listed at https://squiggle.sh/llms.txt, and all of them are one file at https://squiggle.sh/llms-full.txt.
The Screen map shows every screen Squiggle has seen in your app, and marks the ones no saved flow reaches. Start an agentic flow on one of them, write its goal, and it's covered.
See which screens have no flow
Open the Inspector’s Screen map. Screens no flow reaches yet are marked; No flow yet shows only those. Select a screen to see the taps that reach it, the flows that already do, and the testIDs on it.
Start a flow on a screen
Select an uncovered screen and press New agentic flow here. Studio saves a flow that starts on that screen with an agent step, and opens it in Flows. Write the goal in plain English, add the assert that proves it, and run it with your agent. See Agentic flows.
Fill in the map
The map grows as you use the app, run flows and crawl. To visit every screen your app can reach, run a complete crawl:
squiggle crawl --completeWhen it finishes, it lists the screens your app declares that it couldn’t reach. See Record and crawl.