harness-bench— agent skill

Surfaces metaharness-darwin bench <create|verify> — the supporting verb for harness-evolve --bench. Use when you want evolution scored against a fixed corpus (independent of npm test) so champion fitness is comparable across commits or across forks of the same harness.

When to use

• Setting up a new evolution pipeline for a repo whose npm test is

flaky, slow, or undersized — scaffold a deterministic bench suite once, then evolve against it repeatedly.

• CI: bench verify the checked-in suite on every PR that touches it

(cheap; ~5s).

• Forking a harness to a new domain: copy and edit the suite to retarget

the evaluation without losing comparability to the parent.

Algorithm

Implementation: scripts/bench.mjs.

--op create

1. Resolve --repo path; reject if missing.
2. Shell to metaharness-darwin bench create <repo> [--out <suite.json>].
3. Default output path: <repo>/.metaharness/bench/suite.json (chosen by upstream).
4. Suite shape (per upstream): array of { input, expectedOutput, weight } tasks

derived from existing test cases.

--op verify

1. Resolve --suite path; reject if missing.
2. Shell to metaharness-darwin bench verify <suite.json>.
3. Exit 1 if any task malformed (upstream's signal).

Output shape

{
  "success": true,
  "data": {
    "op": "verify",
    "taskCount": 42,
    "wellFormed": true,
    "durationMs": 870
  }
}

Exit codes

Graceful degradation

When @metaharness/darwin is absent, emits the standard {degraded: true, reason: 'metaharness-darwin-not-available'} payload and exits 0.