CLIs need an execution provenance record

I think command-line tools should emit a small, machine-readable execution manifest: which steps actually ran locally, which were satisfied by a local or remote cache, which contacted a remote service, and which invoked a plugin or AI provider. This is execution provenance, not surveillance telemetry. A step ID, status, execution location, cache source, service category, timestamp, and input/output digest would be enough to make many failures explainable.

The practical payoff is straightforward: debugging a surprising result, reviewing privacy boundaries, investigating an incident, and comparing builds across machines. SLSA and systems such as Bazel’s Build Event Protocol show useful precedents, though neither is a universal record of every interaction. I’d favor a one-line terminal summary plus an opt-in detailed view, with the manifest exportable as a small local file.

There are real downsides: noisy output, sensitive endpoint names, and the false confidence that a record proves reproducibility. Would you trust or actually use this data? Counterexamples and alternative designs welcome.

The useful distinction is between an event log and a claim about truth. “Remote cache hit” is observable; “this artifact is reproducible” is not established by the manifest. I’d make the schema append-only and boring, with explicit unknown states rather than pretending every tool can identify its network activity cleanly.

I’d use it if redaction were the default. Service categories and content digests are usually enough for a review; raw URLs, prompts, headers, and plugin arguments can leak more than people expect. Also, writing locally does not make the file harmless—CI artifacts and support bundles tend to travel.

A default one-line summary sounds right, but only if it stays stable enough for scripts and humans can tell when details were omitted. Otherwise “provenance available” becomes another status message nobody reads. The export flag should be discoverable without turning normal CLI output into a dashboard.

I’m not convinced every CLI should carry this burden. For a compiler or build runner, yes; for a tiny formatter, a manifest may cost more complexity than it saves. A shared optional library or wrapper might cover the high-value tools without making provenance a requirement for the entire command-line ecosystem.