CLI
Probierz CLI command reference.
Every public Probierz command has one canonical page. Nested command groups keep their slash-separated route, and every page records the exact invocation, inputs, effects, and fail-closed behavior exposed by the installed Rust binary.
Command tree#
Choose a command below. Group entries document the dispatcher boundary; indented arrow labels are runnable leaf subcommands.
Output#
A command prints its answer as one JSON document unless its page says it prints text. `--text`, accepted by every command, prints the same answer as indented `key: value` lines for a person, each list entry on its own `- ` line; the commands that print text by default take `--json` instead. A refusal is printed on standard error in both forms.
Source snapshot#
This reference is generated from the Rust binaries at https://github.com/wisent-ai/probierz.git revision `f65b755da397f2ec7435eb80ed722aa6017437c1`, not an unpinned latest checkout.
The captured binary reports `probierz 0.1.5` and contains 91 source-bound command invocations. Combined help SHA-256: `4a12655d233812abc29eeebbf3db3065efe3e00288024383f3eef3970d092e94`; binary SHA-256: `d5398832d3ab52c541b1efe979730dea62941ac6197898a20fccfc5d3fca44df`.
Discovery and product status#
- probierz list — List every supported test surface together with its package, tool, scripts, targets, and required environment coordinates.
- probierz apps — List registered products, their targets, and declared journeys.
- probierz app — Load and print one validated Probierz application manifest.
- probierz identity — Compute the exact Probierz harness and application source identity used to bind evidence to code.
- probierz specs — Discover spec files currently present on disk, optionally for one surface.
- probierz describe — Statically extract the `describe` and test titles from a spec file.
- probierz cmd — Print the exact shell command registered for one execution target without running it.
- probierz accessibility — Validate stable accessibility IDs, native selectors, and selector references for a registered application.
- probierz history — Read stability history by run, journey, and test for an application, optionally for one target and one journey, with the newest passing run of that selection.
- probierz dashboard — Project product, version, journey, and evidence state for one application.
- probierz status — Report journey coverage, freshness against a base ref, affected journeys, evidence levels, and pull-request merge eligibility.
- probierz hosts — List local and Stado execution hosts known to Probierz.
- probierz overview — Compose journeys, merge eligibility, first-use journey state, repository violations, and Stado fleet health into one all-product or selected-product view.
- probierz errors — Produce the fast all-product failure view without repository violation scans.
Setup and existing definitions#
- probierz onboarding — Show the first-run journey and optionally adopt definitions from an existing Probierz repository without running them.
- probierz project — Enter the durable project-adoption command group.
- probierz project → adopt — Adopt existing Probierz application manifests and specs into the current project without executing them.
- probierz project → adoptions — List the retained source identities for previously adopted project definitions.
Application capabilities#
- probierz apphook — Run one built-in application setup, broker, or evaluation capability named by a product manifest.
- probierz mcp — Serve the MCP protocol on stdio: the same commands as tools, run in this process, until the client closes its end.
Failure intake#
- probierz failures — Read the Wisent Errors intake index, including counts by service and error code plus the newest envelopes.
- probierz intake — Enter the failure-intake command group.
- probierz intake → serve — Listen for authenticated `wisent-errors` failure envelopes from desktop applications.
- probierz incident — Enter the register of attempts that did not hold.
- probierz incident → record — Record one attempt that did not hold, with the claim in the words it was claimed in.
- probierz incident → list — Read the register, newest first, for the state the caller names.
- probierz incident → show — Read one register entry with its failure envelope and its resolution.
- probierz incident → resolve — Close one register entry by recording what actually repaired it.
- probierz incident → reopen — Open a resolved incident again, saying why its resolution did not hold.
Local execution and analysis#
- probierz check — Check whether one target's toolchain and host prerequisites are ready, and report the owner and fix for anything missing.
- probierz setup — Install the browser or Appium-driver components Probierz owns for one target, then report preflight state.
- probierz run — Execute one target, capture supported evidence, and analyze its report.
- probierz analyze — Parse a completed runner report, inventory its media, and optionally extract a frame montage.
- probierz gif — Render a silent, looping README GIF from one real journey recording and write its provenance sidecar.
- probierz affected — Map a git change or explicit file list to the targets it can affect.
- probierz ci — Select affected targets, run the ready targets, analyze completed runs, and return one change-driven verdict.
- probierz evaluate — Enter the evaluation group: evaluate figures, SEO contracts, talks or explicitly owned HTTP security scope.
- probierz evaluate → figure — Render a reference and candidate figure deterministically, then obtain a rubric-scored vision verdict through the authenticated Stado model router.
- probierz evaluate → seo — Run the full crawl, indexability, structured-data, dual-model content, performance, production-evidence, and signed SEO evaluation contract.
- probierz evaluate → talk — Judge a rendered talk's slides and speaker script against the documents it presents, through the authenticated Stado model router.
- probierz journal — Read retained security progress without contacting any target or model.
- probierz evaluate → security — Execute declared security checks only against an owned application, optionally asking an explicit Brama model to prioritize them and explain results.
- probierz matrix — Plan or execute an application's declared nightly or release condition matrix.
Remote Stado execution#
- probierz remote — Enter Probierz's remote Stado execution command group.
- probierz remote → run — Submit one target to a selected Stado host and return its evidence to the local `test-results` store.
- probierz remote → collect — Read an existing job's state immediately and recover its retained evidence if it is terminal.
- probierz remote → resume — Resume waiting for an existing remote job and recover its original evidence.
- probierz remote → cancel — Cancel the original Stado job and retain its cancellation record, logs, source identity, and available evidence.
- probierz remote → author — Author and verify one product-owned journey on a Stado host, then return its accepted bytes to the exact submitting checkout.
- probierz remote → seo — Execute the complete SEO evaluator on a Stado-selected dedicated host.
Benchmark#
- probierz benchmark — Enter the benchmark group: run our product and its rivals on the same versioned cases, read the recorded runs, and keep the catalog's rivals and roadmap honest.
- probierz benchmark → suites — Read the suites and contenders one product's manifest declares, with each suite's version, hash and case count.
- probierz benchmark → run — Run every case of one suite against the chosen contenders, judge each answer and record the run.
- probierz benchmark → list — Read the recorded runs of one product, newest first.
- probierz benchmark → show — Read one recorded run with every sample.
- probierz benchmark → compare — Compare two runs of the same suite contender by contender.
- probierz benchmark → standing — Read the newest run of one suite case by case: who won, and every case ours did not win.
- probierz benchmark → rivals — Confront the rivals Stado's product catalog names with the contenders and suites the manifest declares.
- probierz benchmark → roadmap — Turn the newest run's lost cases into the product's catalog roadmap, and record the items of cases ours now wins as delivered with the run that won them.
- probierz benchmark → pursue — Hand one case the newest run lost to a Jeden pursuit in our product's checkout, then decide it with a new run Probierz records itself.
- probierz benchmark → author — Enter the authoring group: draft, judge and declare what a benchmark is made of, a suite or one contender's driver, so neither is written by hand for one product.
- probierz benchmark → author → suite — Draft a suite for any product and its rivals, judge it, write it once and declare it.
- probierz benchmark → author → contender — Draft, place, declare and verify the driver of one contender, ours or a rival the catalog names, so no driver is written by hand.
- probierz benchmark → scout — Scout a product that does not exist yet from a Trends topic: the products the topic's observations show, the rivals a new product must beat among them, and the suite that measures it.
- probierz benchmark → adopt — The operator's decision on one scouted brief: Stado creates the product's private repository and preview catalog record, the catalog names its rivals and benchmark, and Probierz drafts its first suite.
- probierz benchmark → cycle — One pass of the whole benchmark loop without the operator, under the policy he wrote once: new markets watched, new products scouted, judged and created, every suite run, losses pursued and failures turned into roadmap work.
- probierz benchmark → schedule — Have Stado run the benchmark loop on a cron, pinned to one host, so nothing waits for anyone to type it.
Evidence, security, and publication#
- probierz secrets — Enter the secrets group: scan an artifact directory for credentials and tokens.
- probierz secrets → scan — Scan an artifact directory for secrets before protected storage or publication.
- probierz audit — Read the evidence access and security audit trail with optional filters.
- probierz protect — Encrypt one run's artifact bundle at rest and optionally remove the plaintext source artifacts.
- probierz restore — Decrypt a protected evidence bundle into an explicit destination.
- probierz retention — Evaluate or apply the configured evidence-retention policy for an application.
- probierz compare — Create a deterministic diff between two recorded runs.
- probierz receipt — Enter the receipt group: sign exact runs into a release receipt, or verify one.
- probierz receipt → create — Validate selected evidence against exact source identities and sign one release receipt.
- probierz receipt → verify — Verify a receipt's payload hash, Ed25519 signature, trust, and optional expected public-key fingerprint.
- probierz publication — Emit an immutable verified first-use publication manifest from a signed receipt and explicit asset catalog. The journey's version, version ID, first-success fact and screen are read from the signed run and checked against the app manifest; this is the one command that writes a probierz-first-use-publication manifest.
Release gates#
- probierz gate — Enter the gate group: the repository pre-push gate and the pull-request and release gates.
- probierz gate → status — Read the installed pull-request and release gate enforcement state for an application.
- probierz gate → prepush — Evaluate the repository's pre-push merge gate, optionally running change-driven CI first.
- probierz gate → install — Install the managed Probierz pre-push gate into a repository while chaining one existing unmanaged hook.
- probierz gate → uninstall — Remove the pre-push gate `gate install` wrote and put back the hook it chained.
- probierz gate → evaluate — Evaluate evidence against an application's pull-request or release policy without enforcing or activating the gate.
- probierz gate → enforce — Enforce an already activated pull-request or release gate against the supplied exact evidence.
- probierz gate → activate — Activate enforcement for one gate mode only after the exact current evidence evaluates green.
- probierz gate → deactivate — Withdraw one activated gate mode, so its policy is evaluated but no longer enforced.