spex

Plugin spex from marketplace claude-my-marketplace in lukaskellerstein/claude-my-marketplace. Commands read from the repo's manifest — 5 days ago.

Skills (12)

• approve

  A person's approval, as a skill — runs spex approve <dir> [stage] at once for the stage that waits (the spec, the draft records with their proof plan, tasks.yaml, or tests.yaml alone), with no question, because typing the command is the approval. Then says what was approved, in a short digest with t

• clarify

  Resolve the open questions of a spex spec, each with options and a recommended answer, and write the answers and the record changes back into requirements.yaml. Use when the user says "clarify", "spex clarify", "answer the spec questions", or when spex check reports REQ-QUESTION-OPEN. With "auto", o

• develop

  Build spex tasks — one task here, or, given only the spec folder, every ready task through fresh developer subagents, wave by wave, with spex done (the task's own tests green) as the only completion. A wrong test is judged by a decider and rewritten by a test-developer subagent, never by the develop

• init

  Set up spex in a repository once — write .spex/config.yaml from a scan of the test runner, scripts and folders (the default driver, a playwright driver when the repo has it, fast checks, the real systems a test may write, the people who are never an actor, how a feature ships), plus the .spex/.gitig

• new

  Shape a raw idea into a spec in rounds — size it (one step or a roadmap of planned steps), then every open question on one screen per round, each with a pick, until the user says done. The only way in; resumes a shaping spec, opens a planned step. With "auto" a decider answers and nobody is asked.

• requirements

  Turn an approved spec folder (made by /spex:new) into requirements.yaml — EARS records with real-valued scenarios, questions, unknowns — make it pass spex check, settle the questions (the decider in auto mode, the user otherwise) and the unknowns, then have a test-analyst subagent write the proof pl

• research

  Decide the open unknowns of a spex spec — library choices, versions, how an existing API behaves — and write research.md plus a one-line decision per unknown into requirements.yaml. Use when the user says "research", "spex research", "resolve the unknowns", or when spex gate plan reports GATE-UNKNOW

• review

  Review a verified spex feature from its diff, read only, for what no test sees — unrequested changes, a broken boundary, a signature that differs from produces, a bug or a hole the tests missed — and route each finding into a fix task, a question, a test rewritten by a test-developer subagent, or th

• ship

  Ship a reviewed spex feature with no click — spex ship flips its approved records to built, retires the records it supersedes, commits when ship.commit says so and merges through ship.cmd (gh, glab, the forge API); then surface every answer given by default, and, for a step of a roadmap, route what

• tasks

  Turn approved spex requirements into specs/NNN-slug/tasks.yaml — approach, boundaries, a design block (modules, entities, decisions) and one task record per agent run with write allowlists, context, signatures and named values, never a person as an actor — make it pass spex check with full coverage,

• update

  After a spex plugin update, bring the repo to the plugin's format — run spex update, which adds the keys the newer plugin needs with their defaults and comments, runs the format migrations (1 → 2 moves every record's verification and every task's verify commands into a tests.yaml per folder) and set

• verify

  Prove a built spex feature as a whole — spex verify runs fast_checks and every script test at this tree, tester subagents perform every session test (a browser through the bundled Playwright MCP, an API, a CLI) and record what they saw, and each failure goes to a decider (test, code or spec?) and th

Subagents (7)

• decider

  Settles the open questions of one spex step from evidence: the design sections the step names, the stub, what the steps before it decided, the code a question points at, and the facts researchers brought back. Returns one verdict per question with its evidence and a confidence, or what it still need

• developer

  Builds one spex task inside its write allowlist, with the task's own tests (from the proof plan, tests.yaml) as the outer loop. The orchestrating /spex:develop starts one per ready task, in parallel for a parallel group, and runs `spex done` itself afterwards. It gets a spec folder and a task id, re

• researcher

  Answers one research question for a spex spec with sources: how others solve it, what a library or tool does and costs, what a standard says, what this repo already has. Returns the answer, the links it opened, a confidence, and what it did not find. /spex:new and /spex:research start one per questi

• reviewer

  Reviews one task's slice of a spex feature diff, read only, for what no test sees: a spec verdict per requirement (met by the evidence, partial, missing, contradicts, cannot-verify), the unrequested changes, and blocking findings with file and line. /spex:review starts one per task when a feature ha

• test-analyst

  Writes the proof plan of a spex feature, tests.yaml: one test per claim — a scenario of a record, or a task's done_when — as a script a driver runs, a shell test, or a session a tester performs when no script can judge. It never wrote the claim and never writes the test code. /spex:requirements star

• test-developer

  Writes the code of one or more script tests of a spex proof plan (tests.yaml), each from its record alone, and proves each red with spex run --test --expect fail before the code exists, or by a mutation when the code is there already. It never reads the source and never edits tests.yaml. /spex:tests

• tester

  Performs one session test of a spex proof plan — the steps a script cannot run: a browser through the bundled Playwright MCP, an API with curl or the forge CLI, a real device — and records what it observed with spex session, at this tree, with screenshots or logs as artifacts. It changes nothing: no

MCP servers (1)

• playwright

  runs: npx -y @playwright/mcp@latest --headless --output-dir .spex/artifacts

Similar plugins