An offline reporter for command results that were already saved by another system. It answers what exit status was recorded, where a salient error-shaped line begins, and what to inspect next. It never executes the command or publishes its output text. A failed exit is not reinterpreted as success when the saved stdout sounds reassuring.
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/command-output-summarizer.mjs --root examples/clean --input input.json
node bin/command-output-summarizer.mjs --root examples/failing --input input.json
node bin/command-output-summarizer.mjs --root examples/incomplete --input input.json
npm run checkRead the result
Those examples exit 0 (pass), 1 (fail) and 2 (incomplete). --help prints usage. Standard output is one JSON report line. Standard error is a fixed human-readable status/count summary; --json suppresses only that summary. Neither stream repeats command, stdout or stderr text. The CLI has no report file mode and never writes to its input.
Scope
Defaults are inclusive: 1,048,576 input bytes, 128 results, 65,536 UTF-16 units and 4,096 lines per output stream, JSON depth 64, JSON structural nodes 100,000, and 2,000 milliseconds of cooperative analysis. Tests pin N and N+1. Library callers may lower the result, stream, line and time limits. The clock cannot interrupt a synchronous parser call; an elapsed limit is checked between records. No arbitrary excerpt is considered provably secret-free, so none is rendered. The tool does not execute commands, diagnose every program-specific error, verify that an exporter told the truth, connect to a host, or repair a failing command. It has no runtime or development dependencies beyond Node.js 22 or newer. npm run check runs the suite under an active network-denial guard.