@pmelab/gtd 58.1.0 → 59.0.0

package/src/workflows/feature/standards.feature

@@ -1,354 +0,0 @@
-@inmem
-Feature: the review learns standards from the human's answers
-
-  A won't-fix tick or an own answer at the review gate — on a note the human
-  added with its own answer boxes — is a candidate standard. A judge asks whether it is reusable; above `standardMinP` it is
-  kept in `.gtd/STANDARDS.md` with no further question, and every later agent
-  turn — lenses and coder alike — is told to follow it. One that contradicts
-  a standard already kept becomes a note in `.gtd/REVIEW.md` the review cannot
-  sign off without: keep the new one, keep the old one, or write the merged
-  wording.
-
-  Background:
-    Given the shipped workflows, as gtd install scaffolds them
-    And an environment variable "GTD_QUALITYREVIEWS" set to ""
-    And a file "src/calc.ts" with:
-      """
-      export const add = (a: number, b: number) => a + b
-      """
-    And the working tree is committed as "feat: add calculator"
-    And a file "TODO.md" with:
-      """
-      # Add subtract
-
-      subtract(a, b) returns the difference.
-      """
-    And step "unwind" changes nothing
-    And step "start-gate.check" changes nothing
-    And step "design.triage" writes ".gtd/REQUIREMENTS.md" with:
-      """
-      # Requirements
-
-      ## Concerns
-
-      ### Subtract
-
-      - `subtract(a, b)` returns `a - b`.
-
-      ## Impact
-
-      - the greeting module
-
-      ## Acceptance Criteria
-
-      - AC-1: subtract(a, b) returns a - b.
-      """
-    And step "design.gate.answer" changes nothing
-    And step "seam.naive" writes "src/calc.test.ts" with:
-      """
-      it("[AC-1] subtracts", () => {})
-      """
-    And step "seam.naive" writes "src/calc.ts" with:
-      """
-      export const add = (a: number, b: number) => a + b
-      export const subtract = (a: number, b: number) => a - b
-      """
-    And step "seam.check" changes nothing
-    And step "implement.applying" runs its script
-    And step "implement.health.check" changes nothing
-    And step "implement.closing" runs its script
-    And step "build.review.reviewing" writes ".gtd/REVIEW.md" with:
-      """
-      # Review: abc1234
-
-      <!-- base: 0000000 -->
-
-      ## Subtract
-
-      - [ ] ./src/calc.ts#2-2 — subtract(a, b)
-      """
-    And step "build.review.carrying" runs its script
-    And step "build.review.closing" runs its script
-    And step "build.review.standards.recording" runs its script
-    And step "build.review.standards.raising" runs its script
-
-  Scenario: a won't-fix the judge calls reusable is kept, and the review signs off
-    Given judge step "build.review.standards.judge" answers:
-      """
-      [{"id":"standard-1","answer":"yes","p":0.9}]
-      """
-    When the process is driven
-    And ".gtd/REVIEW.md" gets appended:
-      """
-      ## subtract has no class
-
-      Calculators read better as classes.
-
-      Tick exactly one:
-
-      - [ ] fix as suggested
-      - [ ] won't fix
-      - [ ] _your answer_
-      """
-    And the text "- [ ] won't fix" in ".gtd/REVIEW.md" is replaced with "- [x] won't fix"
-    And gtd lands
-    And the process is driven
-    Then the process rests at "idle"
-    And ".gtd/STANDARDS.md" contains:
-      """
-      - Won't fix: subtract has no class — Calculators read better as classes.
-      """
-    And the commit subjects end with:
-      """
-      gtd(human): build.review.await-review → build.review.standards.judge
-      gtd(judge): build.review.standards.judge → build.review.standards.recording
-      gtd(check): build.review.standards.recording → build.review.closing
-      gtd(check): build.review.closing → idle
-      """
-
-  Scenario: the next review lists the standards added
-    Given judge step "build.review.standards.judge" answers:
-      """
-      [{"id":"standard-1","answer":"yes","p":0.9}]
-      """
-    And judge step "build.review.triage" answers:
-      """
-      [{"id":"answer-1","answer":"nit","p":0.9}]
-      """
-    And step "build.review.fix.nits.fixing" writes "src/calc.ts" with:
-      """
-      export function add(a: number, b: number) { return a + b }
-      export function subtract(a: number, b: number) { return a - b }
-      """
-    And step "build.health.check" changes nothing
-    When the process is driven
-    And ".gtd/REVIEW.md" gets appended:
-      """
-      ## subtract has no class
-
-      Calculators read better as classes.
-
-      Tick exactly one:
-
-      - [ ] fix as suggested
-      - [ ] won't fix
-      - [ ] _your answer_
-      """
-    And the text "- [ ] _your answer_" in ".gtd/REVIEW.md" is replaced with "- [x] Calculators are plain functions."
-    And gtd lands
-    And the process is driven
-    Then the process rests at "build.review.await-review"
-    And ".gtd/REVIEW.md" contains:
-      """
-      ## Standards added
-
-      Learned from your answers and kept in `.gtd/STANDARDS.md`:
-
-      - Calculators are plain functions. (re: subtract has no class)
-      """
-
-  Scenario: an answer below the threshold is not kept
-    Given judge step "build.review.standards.judge" answers:
-      """
-      [{"id":"standard-1","answer":"yes","p":0.6}]
-      """
-    When the process is driven
-    And ".gtd/REVIEW.md" gets appended:
-      """
-      ## subtract has no class
-
-      Calculators read better as classes.
-
-      Tick exactly one:
-
-      - [ ] fix as suggested
-      - [ ] won't fix
-      - [ ] _your answer_
-      """
-    And the text "- [ ] won't fix" in ".gtd/REVIEW.md" is replaced with "- [x] won't fix"
-    And gtd lands
-    And the process is driven
-    Then the process rests at "idle"
-    And ".gtd/STANDARDS.md" does not exist
-
-  Scenario: a kept standard reaches the lenses and the coder of a later process
-    Given a file ".gtd/STANDARDS.md" with:
-      """
-      # Standards
-
-      - Won't fix: subtract has no class — Calculators read better as classes.
-      """
-    And the working tree is committed as "chore: keep a standard"
-    And a file "TODO.md" with:
-      """
-      # Add subtract
-
-      Please add subtract(a, b).
-      """
-    And step "seam.naive" is prompted with "- Won't fix: subtract has no class"
-    And step "build.quality.correctness.reviewing" is prompted with "- Won't fix: subtract has no class"
-    When the process is driven
-    Then the process rests at "build.review.await-review"
-
-  Scenario: a contradicting standard blocks sign-off until the human picks one
-    Given a file ".gtd/STANDARDS.md" with:
-      """
-      # Standards
-
-      - Calculators are classes.
-      """
-    And the working tree is committed as "chore: keep a standard"
-    And a file "TODO.md" with:
-      """
-      # Add subtract
-
-      Please add subtract(a, b).
-      """
-    And judge step "build.review.standards.judge" answers:
-      """
-      [{"id":"standard-1","answer":"yes","p":0.9},{"id":"standard-1.contradicts-1","answer":"yes","p":0.8}]
-      """
-    And judge step "build.review.triage" answers:
-      """
-      [{"id":"answer-1","answer":"praise","p":0.9}]
-      """
-    When the process is driven
-    And ".gtd/REVIEW.md" gets appended:
-      """
-      ## subtract has no class
-
-      Calculators read better as classes.
-
-      Tick exactly one:
-
-      - [ ] fix as suggested
-      - [ ] won't fix
-      - [ ] _your answer_
-      """
-    And the text "- [ ] _your answer_" in ".gtd/REVIEW.md" is replaced with "- [x] Calculators are plain functions."
-    And gtd lands
-    And the process is driven
-    Then the process rests at "build.review.await-review"
-    And ".gtd/STANDARDS.md" contains:
-      """
-      - Calculators are classes.
-
-      ## Undecided
-
-      - Calculators are plain functions. (re: subtract has no class)
-        - contradicts: Calculators are classes.
-      """
-    And ".gtd/REVIEW.md" contains "## Standard conflict "
-    And ".gtd/REVIEW.md" contains:
-      """
-      New: Calculators are plain functions. (re: subtract has no class)
-
-      Old:
-
-      - Calculators are classes.
-
-      Tick exactly one:
-
-      - [ ] keep the new one (drop the old)
-      - [ ] keep the old one (discard the new)
-      - [ ] _your answer_
-      """
-    And ".gtd/REVIEW.md" does not contain "## subtract has no class"
-    And landing is refused with "answer-completeness: 1 note(s) in .gtd/REVIEW.md not answered"
-    When the text "- [ ] keep the new one (drop the old)" in ".gtd/REVIEW.md" is replaced with "- [x] keep the new one (drop the old)"
-    And gtd lands
-    And the process is driven
-    Then the process rests at "idle"
-    And ".gtd/STANDARDS.md" contains "- Calculators are plain functions. (re: subtract has no class)"
-    And ".gtd/STANDARDS.md" does not contain "Calculators are classes."
-    And ".gtd/STANDARDS.md" does not contain "Undecided"
-
-  Scenario: keeping the old standard discards the new one
-    Given a file ".gtd/STANDARDS.md" with:
-      """
-      # Standards
-
-      - Calculators are classes.
-      """
-    And the working tree is committed as "chore: keep a standard"
-    And a file "TODO.md" with:
-      """
-      # Add subtract
-
-      Please add subtract(a, b).
-      """
-    And judge step "build.review.standards.judge" answers:
-      """
-      [{"id":"standard-1","answer":"yes","p":0.9},{"id":"standard-1.contradicts-1","answer":"yes","p":0.8}]
-      """
-    When the process is driven
-    And ".gtd/REVIEW.md" gets appended:
-      """
-      ## subtract has no class
-
-      Calculators read better as classes.
-
-      Tick exactly one:
-
-      - [ ] fix as suggested
-      - [ ] won't fix
-      - [ ] _your answer_
-      """
-    And the text "- [ ] won't fix" in ".gtd/REVIEW.md" is replaced with "- [x] won't fix"
-    And gtd lands
-    And the process is driven
-    And the text "- [ ] keep the old one (discard the new)" in ".gtd/REVIEW.md" is replaced with "- [x] keep the old one (discard the new)"
-    And gtd lands
-    And the process is driven
-    Then the process rests at "idle"
-    And ".gtd/STANDARDS.md" contains:
-      """
-      # Standards
-
-      Kept from your review answers for every agent turn. One standard per bullet.
-
-      - Calculators are classes.
-      """
-    And ".gtd/STANDARDS.md" does not contain "Won't fix"
-
-  Scenario: an own answer to a conflict replaces both with its wording
-    Given a file ".gtd/STANDARDS.md" with:
-      """
-      # Standards
-
-      - Calculators are classes.
-      """
-    And the working tree is committed as "chore: keep a standard"
-    And a file "TODO.md" with:
-      """
-      # Add subtract
-
-      Please add subtract(a, b).
-      """
-    And judge step "build.review.standards.judge" answers:
-      """
-      [{"id":"standard-1","answer":"yes","p":0.9},{"id":"standard-1.contradicts-1","answer":"yes","p":0.8}]
-      """
-    When the process is driven
-    And ".gtd/REVIEW.md" gets appended:
-      """
-      ## subtract has no class
-
-      Calculators read better as classes.
-
-      Tick exactly one:
-
-      - [ ] fix as suggested
-      - [ ] won't fix
-      - [ ] _your answer_
-      """
-    And the text "- [ ] won't fix" in ".gtd/REVIEW.md" is replaced with "- [x] won't fix"
-    And gtd lands
-    And the process is driven
-    And the text "- [ ] _your answer_" in ".gtd/REVIEW.md" is replaced with "- [x] Calculators with state are classes."
-    And gtd lands
-    And the process is driven
-    Then the process rests at "idle"
-    And ".gtd/STANDARDS.md" contains "- Calculators with state are classes."
-    And ".gtd/STANDARDS.md" does not contain "- Calculators are classes."
-    And ".gtd/STANDARDS.md" does not contain "Won't fix"

package/src/workflows/feature/steps.test.ts

@@ -1,65 +0,0 @@
-import { readFileSync } from "node:fs"
-import { describe, expect, it } from "vitest"
-import type { StepRequest } from "../../flows/index.js"
-import { captureStep } from "../shared/text.fixture.js"
-import * as steps from "./steps.js"
-
-const capture = (
-  fn: () => Promise<void>,
-  scope: string,
-  skills?: Readonly<Record<string, readonly string[]>>,
-): Promise<StepRequest> => captureStep(fn, { scope, ...(skills !== undefined ? { skills } : {}) })
-
-const agentPrompt = (request: StepRequest): string => {
-  if (request.kind !== "agent") throw new Error(`expected an agent step, got ${request.kind}`)
-  return request.prompt
-}
-
-describe("the feature workflow's own steps declare skills — a step's rendered prompt carries its bundled skills.ts preamble", () => {
-  it("triage carries design.triage's bundled skills", async () => {
-    const prompt = agentPrompt(await capture(() => steps.triage("base"), "design"))
-    expect(prompt).toContain("spec-driven-development, planning-and-task-breakdown")
-  })
-
-  it("collecting carries build.review.collecting's bundled skills", async () => {
-    const prompt = agentPrompt(await capture(() => steps.collecting("capture"), "build"))
-    expect(prompt).toContain("code-review-and-quality")
-  })
-
-  it("collecting's prompt is the pinned feature fixture", async () => {
-    const prompt = agentPrompt(await capture(() => steps.collecting("capture-text"), "build"))
-    expect(prompt).toContain(
-      readFileSync(
-        new URL("../shared/__fixtures__/feature-collecting-prompt.txt", import.meta.url),
-        "utf8",
-      ),
-    )
-  })
-
-  it("refactorCollecting renders the open variant for the reply round, the closed one otherwise, whatever .gtd/REVIEW.md holds", async () => {
-    const at = (round: "reply" | "edits", present: boolean): Promise<StepRequest> =>
-      captureStep(() => steps.refactorCollecting("capture", round, ["src/a.ts"]), {
-        scope: "refactor.build",
-        read: (path) => (present && path === ".gtd/REVIEW.md" ? "# Review" : undefined),
-      })
-    const open = agentPrompt(await at("reply", false))
-    const closed = agentPrompt(await at("edits", true))
-    expect(open).toContain("exactly one `- A:` reply")
-    expect(open).not.toContain("already closed")
-    expect(closed).toContain("already closed")
-    expect(closed).not.toContain("exactly one `- A:` reply")
-  })
-})
-
-describe("the feature workflow's planning steps run on the planner model", () => {
-  it("design triage defaults to smart and follows plannerModel", async () => {
-    const plain = await captureStep(() => steps.triage("base"), { scope: "design" })
-    const set = await captureStep(() => steps.triage("base"), {
-      scope: "design",
-      env: { plannerModel: "opus" },
-    })
-    if (plain.kind !== "agent" || set.kind !== "agent") throw new Error("unreachable")
-    expect(plain.options.model).toBe("smart")
-    expect(set.options.model).toBe("opus")
-  })
-})

package/src/workflows/feature/steps.ts

@@ -1,45 +0,0 @@
-import { human } from "../../flows/index.js"
-import type { CollectRound } from "../../workflow-kit/index.js"
-import * as t from "../shared/text.js"
-import { riskTierOn } from "./tier.js"
-import { planner, REFACTOR, REQUIREMENTS } from "../shared/steps.js"
-
-// Kept on the published feature entry.
-export { answerReviewQuestions, collecting, fixNits } from "../shared/steps.js"
-
-/** Turn the sketch since `base` into `.gtd/REQUIREMENTS.md`. */
-export const triage = (base: string): Promise<void> =>
-  t.agentWithSkills("triage", t.designTriagePrompt(base, riskTierOn()), {
-    label: "Triaging the change",
-    file: REQUIREMENTS,
-    mode: "qa",
-    model: planner(),
-    system: t.designSystem(),
-  })
-
-export const answerProductQuestions = (): Promise<void> =>
-  human("gate.answer", {
-    message: t.designGateAnswerMessage(riskTierOn()),
-    label: "Awaiting your product answers",
-    file: REQUIREMENTS,
-    mode: "qa",
-    acceptClean: true,
-  })
-
-/** Turn the refactor review round `capture` describes into `.gtd/REFACTOR.md`, or change nothing. */
-export const refactorCollecting = (
-  capture: string,
-  round: CollectRound,
-  strays: readonly string[],
-): Promise<void> =>
-  t.agentWithSkills(
-    "review.collecting",
-    t.buildReviewCollectingPrompt(capture, { strays, round }),
-    {
-      label: "Collecting your refactor feedback",
-      file: REFACTOR,
-      model: planner(),
-      system: t.reviewerSystem(),
-      allowEmpty: true,
-    },
-  )

package/src/workflows/feature/tier-pin.feature

@@ -1,66 +0,0 @@
-@inmem
-Feature: a process keeps the risk tier setting it started with when GTD_FLAG_RISK_TIER changes
-
-  The flag is recorded in the opening commit, so flipping it mid-process never
-  changes whether the process asks for and records a risk tier.
-
-  Background:
-    Given the shipped workflows, as gtd install scaffolds them
-    And an environment variable "GTD_QUALITYREVIEWS" set to ""
-    And a file "src/calc.ts" with:
-      """
-      export const add = (a: number, b: number) => a + b
-      """
-    And the working tree is committed as "feat: add calculator"
-    And a file "TODO.md" with:
-      """
-      # Add subtract
-
-      subtract(a, b) returns the difference.
-      """
-    And step "unwind" changes nothing
-    And step "start-gate.check" changes nothing
-    And step "design.triage" writes ".gtd/REQUIREMENTS.md" with:
-      """
-      # Requirements
-
-      ## Concerns
-
-      ### Subtract
-
-      - `subtract(a, b)` returns `a - b`.
-
-      ## Risk Tier
-
-      low — a new pure function
-
-      ## Impact
-
-      - the greeting module
-
-      ## Acceptance Criteria
-
-      - AC-1: the change behaves as described.
-      """
-
-  Scenario: a process started with the flag off records no tier after the flag turns on
-    Given the ambient environment variable "GTD_FLAG_RISK_TIER" is "0"
-    When the process is driven
-    Then the process rests at "design.gate.answer"
-    Given the ambient environment variable "GTD_FLAG_RISK_TIER" is "1"
-    And step "design.gate.answer" changes nothing
-    When the process is driven
-    Then the process rests at "seam.naive"
-    And the commit "gtd(human): idle → unwind" has a body containing "Gtd-Var: flag_risk_tier=0"
-    And no commit body contains "Gtd-Mark:"
-
-  Scenario: a process started with the flag on records its tier after the flag turns off
-    Given the ambient environment variable "GTD_FLAG_RISK_TIER" is "1"
-    When the process is driven
-    Then the process rests at "design.gate.answer"
-    Given the ambient environment variable "GTD_FLAG_RISK_TIER" is "0"
-    And step "design.gate.answer" changes nothing
-    When the process is driven
-    Then the process rests at "seam.naive"
-    And the commit "gtd(human): idle → unwind" has a body containing "Gtd-Var: flag_risk_tier=1"
-    And the last commit body contains "Gtd-Mark: risk-tier=low"