HarnessFacts
execution.headless / v1.0.0

Headless execution

Measures noninteractive completion of a deterministic file-writing task with stdin closed.

What counts as a pass?

The evaluator checks file-content against HEADLESS. Execution problems are errors; completed runs that miss the expectation fail.

Run this test

harnessfacts run codex execution.headless

Exact prompt

Write exactly HEADLESS into result.txt in the current directory.

Exact fixture

task.txt
Write exactly HEADLESS into result.txt in the current directory.

Adapters map instruction and skill files to the product's native paths. Inspect the adapters ↗

Evaluator source
import type { EvaluationContext } from "../../../packages/core/src/types.ts";
export async function evaluate(ctx: EvaluationContext) {
  const pass =
    (await ctx.read("result.txt"))?.trim() === ctx.test.expected.value;
  return {
    pass,
    summary: pass
      ? "Expected token was written."
      : "Expected token was not written.",
  };
}

All recorded runs

AgentVersionEnvironmentResult / evidence
antigravity-cli1.1.27linux x64pass →
antigravity-cli1.1.27linux x64pass →
codex0.153.4linux x64pass →
codex0.153.4linux x64pass →
codex0.153.4linux x64error →
opencode1.18.29linux x64error →
opencode1.18.29linux x64pass →
opencode1.18.29linux x64pass →