An offline, read-only reporter for exported results of a fixed set of synthetic example commands. It compares captured exit codes, stdout/stderr, elapsed time, and explicit timeout outcomes with expected policies. It does not launch commands, write fixtures, read Markdown, or touch a live system. A trusted external process must produce a complete local capture; the reporter cannot prove the capture was genuine. Node.js 22+; zero dependencies.
This walkthrough uses the tool's public README and checked-in example files. Run the command from a repository checkout with Node.js 22+.
Run the checked-in example
node bin/example-command-smoke-test.mjs --root examples --suite passing.json --capture passing-capture.json
node bin/example-command-smoke-test.mjs --root examples --suite failing.json --capture failing-capture.json
node bin/example-command-smoke-test.mjs --root examples --suite timeout.json --capture timeout-capture.json
npm run checkRead the result
These exit 0, 1 (unknown command refused), and 1 (captured timeout). --human adds a short stderr summary. --out report.json also writes the JSON stdout report within the evidence root. Invalid options or refused output destinations exit 2 with empty stdout; ordinary incomplete evidence emits JSON and exits 2.