Test audit

Adapted from openclaw's test-audit skill (MIT, OpenClaw Foundation; see LICENSE).

Judge the result by how much the remaining tests can be trusted, never by how many were deleted. A test is worth keeping when it protects observable behavior, a credible regression, or an independent contract: a public API, protocol, config, migration, storage, security, platform, default, generated, package, release or architecture contract. Static or slow is never a reason to delete.

Scope

Take the user's scope, or the files they name. Read the root and scoped AGENTS.md first; their Tests rule is the bar for every test this audit keeps, rewrites or moves. Leave out skills and code installed from another repository.

Discover, read-only

1. For a broad scope, split it into read-only lanes by owner area and run them in parallel.
2. For each candidate, read the whole test, its production owner, the owner's callers, the overlapping tests, CI routing and history. When a test claims behavior a dependency provides, read that dependency.
3. Prove what a test catches by mutating a scratch copy of its owner and running the test against it, never by editing the checkout.
4. Prefer a few high-confidence candidates over a long speculative list. Report the evidence before editing.

Junk patterns