Agent Friendly Code
← all checks

Test suites: why AI coding agents need tests they can find

This check looks for tests an agent could run: a test directory, or files named the way test files are named in each language.

How many repos pass

Of the 371 repos on the leaderboard, 193 pass in full (52%), 148 pass in part (40%) and 30 fail (8%).

Why agents need it

Tests are how an agent learns whether its change worked. An agent with a test suite can edit, run and correct itself. An agent without one hands back code nobody has run. Claude Code's profile weighs tests as heavily as AGENTS.md.

How we score it

  • Full pass: a tests/, test/, __tests__/, spec/, specs/ or src/test/ directory.
  • 70%: no such directory, but test-named files turn up within three folder levels, such as *.test.ts, test_*.py, *_test.go or *Test.java.
  • Fail: neither. We check that tests exist, not that they pass.

The methodology has the formula that turns checks into a score.

How to pass

Keep tests in a conventional directory, and write the command that runs them in AGENTS.md. Tests that sit next to the source score 70%, because an agent has to search for them.

Which agents count it most

Pass rate by language

Top-scoring repos that pass

Popular repos that fail it