@delete101020/aidlc 4.0.0

package/dist/templates/speckit/artifacts/analyze.md

@@ -0,0 +1,48 @@
+# Analysis — [Feature Title]
+
+**Epic ID:** `$EPIC_ID`
+**Author:** QA
+**Status:** Draft
+**Created:** `$DATE`
+**Cross-checks:** `SPEC.md` ↔ `PLAN.md` ↔ `TASKS.md`
+
+---
+
+## 1. Coverage Matrix
+
+| Requirement | PLAN section | TASKS task id(s) | Covered? |
+|-------------|--------------|------------------|----------|
+| `$EPIC_ID-FR01` | | | ✅ / ❌ |
+| `$EPIC_ID-NFR01` | | | |
+
+## 2. Gaps
+
+> Requirements with no plan coverage or no task; acceptance criteria no task satisfies.
+
+- …
+
+## 3. Orphans (scope creep)
+
+> Plan decisions or tasks with no upstream requirement.
+
+- …
+
+## 4. Contradictions
+
+> Where any two of spec / plan / tasks disagree.
+
+- …
+
+## 5. Constitution Check
+
+- …
+
+## 6. Verdict
+
+**GO** / **NO-GO**
+
+> If NO-GO, list each blocking issue and which artifact owns the fix.
+
+| Blocking issue | Owner artifact |
+|----------------|----------------|
+|                |                |

package/dist/templates/speckit/artifacts/clarify.md

@@ -0,0 +1,64 @@
+# Spec — [Feature Title]
+
+**Epic ID:** `$EPIC_ID`
+**Author:** Analyst
+**Status:** Draft
+**Created:** `$DATE`
+
+> Describe **what** and **why**, never **how**. No libraries, schemas, or
+> frameworks in this document — those belong in PLAN.md.
+
+---
+
+## 1. Overview
+
+> One paragraph: what this feature is and the user problem it solves.
+
+## 2. User Scenarios
+
+### Primary flow
+- **Given** …
+- **When** …
+- **Then** …
+
+### Edge & error paths
+| Situation | Expected behavior |
+|-----------|-------------------|
+| Dependency unavailable | |
+| Permission / access denied | |
+| Session expired mid-flow | |
+| Empty state | |
+| Boundary input | |
+| Interruption / recovery | |
+
+## 3. Functional Requirements
+
+| ID | Requirement | Priority (MoSCoW) |
+|----|-------------|-------------------|
+| `$EPIC_ID-FR01` | | Must |
+| `$EPIC_ID-FR02` | | Should |
+
+## 4. Non-Functional Requirements
+
+| ID | Category | Requirement (quantified) |
+|----|----------|--------------------------|
+| `$EPIC_ID-NFR01` | Performance | p95 < … ms |
+| `$EPIC_ID-NFR02` | Security / privacy | |
+| `$EPIC_ID-NFR03` | Accessibility | WCAG 2.1 AA |
+
+## 5. Acceptance Criteria
+
+| ID | Given / When / Then |
+|----|---------------------|
+| `$EPIC_ID-AC01` | Given … When … Then … |
+| `$EPIC_ID-AC02` | |
+
+## 6. Out of Scope
+
+- …
+
+## 7. Clarifications
+
+> Populated by the Clarify phase. Each entry: question → resolved answer → decided by.
+
+_None yet — open items are marked `[NEEDS CLARIFICATION: …]` inline above._

package/dist/templates/speckit/artifacts/plan.md

@@ -0,0 +1,57 @@
+# Plan — [Feature Title]
+
+**Epic ID:** `$EPIC_ID`
+**Author:** Tech Lead
+**Status:** Draft
+**Created:** `$DATE`
+**Derived from:** `SPEC.md`
+
+---
+
+## 1. Approach
+
+> The overall technical strategy in a paragraph. Why this shape over the alternatives.
+
+## 2. Architecture & Components
+
+| Component | New / Modified | Responsibility | Interacts with |
+|-----------|----------------|----------------|----------------|
+|           |                |                |                |
+
+## 3. Data Model
+
+> Entities, fields, relationships, migrations. Omit if not applicable.
+
+## 4. Contracts
+
+> APIs / interfaces / events — signature, inputs, outputs, error behavior.
+
+| Contract | Input | Output | Errors |
+|----------|-------|--------|--------|
+|          |       |        |        |
+
+## 5. Constitution Check
+
+> Every workspace-standard principle: satisfied / N/A / tension (+ mitigation).
+
+| Principle | Status | Note |
+|-----------|--------|------|
+|           |        |      |
+
+## 6. File Impact
+
+| Path | New / Modified / Removed | Note |
+|------|--------------------------|------|
+|      |                          |      |
+
+## 7. Requirement Traceability
+
+| Requirement | Plan section that implements it |
+|-------------|--------------------------------|
+| `$EPIC_ID-FR01` | |
+| `$EPIC_ID-NFR01` | |
+
+## 8. Risks & Open Decisions
+
+- **Risk:** … — mitigation …
+- **Decision deferred to implementation:** …

package/dist/templates/speckit/artifacts/specify.md

@@ -0,0 +1,64 @@
+# Spec — [Feature Title]
+
+**Epic ID:** `$EPIC_ID`
+**Author:** Analyst
+**Status:** Draft
+**Created:** `$DATE`
+
+> Describe **what** and **why**, never **how**. No libraries, schemas, or
+> frameworks in this document — those belong in PLAN.md.
+
+---
+
+## 1. Overview
+
+> One paragraph: what this feature is and the user problem it solves.
+
+## 2. User Scenarios
+
+### Primary flow
+- **Given** …
+- **When** …
+- **Then** …
+
+### Edge & error paths
+| Situation | Expected behavior |
+|-----------|-------------------|
+| Dependency unavailable | |
+| Permission / access denied | |
+| Session expired mid-flow | |
+| Empty state | |
+| Boundary input | |
+| Interruption / recovery | |
+
+## 3. Functional Requirements
+
+| ID | Requirement | Priority (MoSCoW) |
+|----|-------------|-------------------|
+| `$EPIC_ID-FR01` | | Must |
+| `$EPIC_ID-FR02` | | Should |
+
+## 4. Non-Functional Requirements
+
+| ID | Category | Requirement (quantified) |
+|----|----------|--------------------------|
+| `$EPIC_ID-NFR01` | Performance | p95 < … ms |
+| `$EPIC_ID-NFR02` | Security / privacy | |
+| `$EPIC_ID-NFR03` | Accessibility | WCAG 2.1 AA |
+
+## 5. Acceptance Criteria
+
+| ID | Given / When / Then |
+|----|---------------------|
+| `$EPIC_ID-AC01` | Given … When … Then … |
+| `$EPIC_ID-AC02` | |
+
+## 6. Out of Scope
+
+- …
+
+## 7. Clarifications
+
+> Populated by the Clarify phase. Each entry: question → resolved answer → decided by.
+
+_None yet — open items are marked `[NEEDS CLARIFICATION: …]` inline above._

package/dist/templates/speckit/artifacts/tasks.md

@@ -0,0 +1,33 @@
+# Tasks — [Feature Title]
+
+**Epic ID:** `$EPIC_ID`
+**Author:** Tech Lead
+**Status:** Draft
+**Created:** `$DATE`
+**Derived from:** `PLAN.md`
+
+---
+
+## Task List
+
+> Ordered by dependency. Check each off during Implement.
+
+### `$EPIC_ID-T01` — [task name]
+- [ ] **What:** …
+- **Depends on:** _(none)_
+- **Parallel-safe:** yes / no
+- **Implements:** `$EPIC_ID-FR01` / PLAN §…
+- **Done when:** … (incl. the test to write)
+
+### `$EPIC_ID-T02` — [task name]
+- [ ] **What:** …
+- **Depends on:** `$EPIC_ID-T01`
+- **Parallel-safe:** no
+- **Implements:** `$EPIC_ID-FR02`
+- **Done when:** …
+
+## Coverage self-check
+
+- [ ] Every plan component has ≥1 task
+- [ ] Every task references a requirement or plan section
+- [ ] Dependencies form a valid order (no cycles)

package/dist/templates/speckit/skills/analyze.md

@@ -0,0 +1,50 @@
+---
+name: analyze
+description: Cross-check spec ↔ plan ↔ tasks for coverage, gaps, and contradictions before implementation, and issue a go/no-go verdict (ANALYSIS.md).
+argument-hint: "<{{EPIC_PREFIX}}-XXXX>"
+---
+
+# Analyze for Epic $0
+
+You are the **QA** agent.
+Load your full persona from `.claude/agents/aidlc-speckit-qa.md` before starting.
+
+## Step 0: Pipeline Gate Check
+Read and execute `.claude/skills/_gate-check.md`. This skill = phase `analyze`, epic = `$0`. If gate fails → STOP.
+
+## Steps
+
+1. Read `docs/epics/$0/artifacts/SPEC.md`, `PLAN.md`, and `TASKS.md`.
+2. Build a coverage matrix: requirement id → plan section → task id(s).
+3. Identify gaps, orphans, and contradictions.
+4. Fill `docs/epics/$0/artifacts/ANALYSIS.md` and issue a verdict.
+
+## ANALYSIS Contents
+
+### Coverage Matrix
+- Table: `$0-FR/NFR id` → PLAN.md section → TASKS.md task id(s). One row per requirement.
+
+### Gaps
+- Requirements with no plan coverage or no task.
+- Acceptance criteria not satisfiable by any task.
+
+### Orphans (scope creep)
+- Plan decisions or tasks with no upstream requirement.
+
+### Contradictions
+- Where any two of spec / plan / tasks disagree.
+
+### Constitution Check
+- Any element that violates a workspace-standard principle.
+
+### Verdict
+- **GO** or **NO-GO**. If NO-GO, list blocking issues and which artifact owns each fix.
+
+## Rules
+
+- 100% of requirements must trace to both plan and tasks for a GO.
+- Do not wave through a NO-GO — route it back to Analyst (spec) or Tech Lead (plan/tasks).
+
+## Output
+
+Write the analysis to `docs/epics/$0/artifacts/ANALYSIS.md`.

package/dist/templates/speckit/skills/clarify.md

@@ -0,0 +1,31 @@
+---
+name: clarify
+description: Surface and resolve the underspecified areas of a spec through structured questions, then fold the answers back into SPEC.md as a Clarifications section.
+argument-hint: "<{{EPIC_PREFIX}}-XXXX>"
+---
+
+# Clarify for Epic $0
+
+You are the **Analyst** agent.
+Load your full persona from `.claude/agents/aidlc-analyst.md` before starting.
+
+## Step 0: Pipeline Gate Check
+Read and execute `.claude/skills/_gate-check.md`. This skill = phase `clarify`, epic = `$0`. If gate fails → STOP.
+
+## Steps
+
+1. Read `docs/epics/$0/artifacts/SPEC.md`.
+2. Collect every `[NEEDS CLARIFICATION: …]` marker plus any requirement that admits more than one reasonable reading.
+3. For each, pose a **closed** question with 2–4 concrete candidate answers (not open-ended). Prefer options the user can pick.
+4. Present the questions to the user and capture their answers. If running non-interactively, choose the most defensible option and mark it as an assumption.
+5. Fold the resolutions back into SPEC.md: update the affected requirements AND append a `## Clarifications` section recording each Q → chosen A (and who decided).
+
+## Rules
+
+- Never leave a `[NEEDS CLARIFICATION]` marker unresolved after this phase — either answered or explicitly deferred with rationale.
+- A clarification that changes scope must update the requirement and its acceptance criteria, not just the log.
+- Keep questions closed and decision-ready; avoid "what do you want here?".
+
+## Output
+
+Update `docs/epics/$0/artifacts/SPEC.md` in place (requirements + a `## Clarifications` section).

package/dist/templates/speckit/skills/plan.md

@@ -0,0 +1,56 @@
+---
+name: plan
+description: Derive the technical implementation plan (PLAN.md) from a clarified spec — architecture, data model, contracts, tech choices — honoring the project constitution and the existing codebase.
+argument-hint: "<{{EPIC_PREFIX}}-XXXX>"
+---
+
+# Plan for Epic $0
+
+You are the **Tech Lead** agent.
+Load your full persona from `.claude/agents/aidlc-speckit-tech-lead.md` before starting.
+
+## Step 0: Pipeline Gate Check
+Read and execute `.claude/skills/_gate-check.md`. This skill = phase `plan`, epic = `$0`. If gate fails → STOP.
+
+## Steps
+
+1. Read `docs/epics/$0/artifacts/SPEC.md` (must be clarified — no open `[NEEDS CLARIFICATION]` markers). If any remain, STOP and send back to Clarify.
+2. Read the active workspace SDLC standard (the constitution) and treat its principles as constraints.
+3. Read the existing codebase around the affected areas (use ast-graph MCP tools for structure/blast-radius when available).
+4. Fill the plan template at `docs/epics/$0/artifacts/PLAN.md`.
+
+## PLAN Contents
+
+### Approach
+- The overall technical strategy, in a paragraph. Why this shape.
+
+### Architecture & Components
+- Modules/components introduced or changed. How they interact.
+
+### Data Model
+- Entities, fields, relationships, migrations (as applicable).
+
+### Contracts
+- APIs / interfaces / events — signatures, inputs, outputs, error behavior.
+
+### Constitution Check
+- For each constitutional principle: satisfied / N/A / tension (with mitigation).
+
+### File Impact
+- Table: path → new / modified / removed → note.
+
+### Requirement Traceability
+- Table: `$0-FR/NFR id` → plan section that implements it. Every requirement must appear.
+
+### Risks & Open Decisions
+- Technical risks, alternatives considered, decisions deferred to implementation.
+
+## Rules
+
+- Every design decision traces to a requirement; every requirement is covered.
+- Reuse existing patterns unless a requirement forces a new one.
+- Do not re-open the "what" — if the spec is wrong or vague, bounce it back, don't paper over it.
+
+## Output
+
+Write the completed plan to `docs/epics/$0/artifacts/PLAN.md`.

package/dist/templates/speckit/skills/specify.md

@@ -0,0 +1,52 @@
+---
+name: specify
+description: Turn a feature description into a structured, testable specification (SPEC.md) — user scenarios, functional and non-functional requirements, measurable acceptance criteria. No implementation detail.
+argument-hint: "<{{EPIC_PREFIX}}-XXXX> [feature description]"
+---
+
+# Specify for Epic $0
+
+You are the **Analyst** agent — a spec-driven-development practitioner.
+Load your full persona from `.claude/agents/aidlc-analyst.md` before starting.
+
+## Step 0: Pipeline Gate Check
+Read and execute `.claude/skills/_gate-check.md`. This skill = phase `specify`, epic = `$0`. If gate fails → STOP.
+
+## Steps
+
+1. Read the epic doc at `docs/epics/$0/$0.md` for intent, scope, and any linked ticket/design.
+2. Read the active workspace SDLC standard (the constitution) so the spec never contradicts a project principle.
+3. Read relevant existing product/domain docs so the spec is consistent with what already ships.
+4. Fill the spec template at `docs/epics/$0/artifacts/SPEC.md`.
+
+## SPEC Contents
+
+### Overview
+- One-paragraph summary of what this feature is and the user problem it solves.
+
+### User Scenarios
+- **Primary flow** — given/when/then from the user's perspective.
+- **Edge & error paths** — dependency down, permission denied, session expired, empty state, boundary inputs, interruption/recovery.
+
+### Functional Requirements
+- Ids `$0-FR01`, `$0-FR02`, … — one testable behavior each. Mark MoSCoW priority.
+
+### Non-Functional Requirements
+- Ids `$0-NFR01`, … covering performance, reliability, security/privacy, accessibility, compatibility, observability as applicable. Quantify.
+
+### Acceptance Criteria
+- Given/When/Then, ids `$0-AC01`, … — measurable, one behavior each, error states included.
+
+### Out of Scope
+- What this feature explicitly does NOT do.
+
+## Rules
+
+- **No implementation detail** — no libraries, schemas, frameworks. Describe *what*, never *how*.
+- Every requirement is testable and carries a stable id.
+- Where you cannot resolve an ambiguity, write `[NEEDS CLARIFICATION: <question>]` rather than guessing — the Clarify phase will resolve it.
+- Quantify success ("> 95%", "< 200ms"), never "should work well".
+
+## Output
+
+Write the completed spec to `docs/epics/$0/artifacts/SPEC.md`.

package/dist/templates/speckit/skills/speckit-implement.md

@@ -0,0 +1,33 @@
+---
+name: speckit-implement
+description: Execute the approved task list on a feature branch, keeping every change traceable to a task, and open a PR when the list is complete.
+argument-hint: "<{{EPIC_PREFIX}}-XXXX>"
+---
+
+# Implement for Epic $0
+
+You are the **Developer** agent.
+Load your full persona from `.claude/agents/aidlc-speckit-developer.md` before starting.
+
+## Step 0: Pipeline Gate Check
+Read and execute `.claude/skills/_gate-check.md`. This skill = phase `implement`, epic = `$0`. If gate fails → STOP.
+
+## Steps
+
+1. Read `docs/epics/$0/artifacts/SPEC.md`, `PLAN.md`, `TASKS.md`, and `ANALYSIS.md` (must be GO — if NO-GO, STOP).
+2. Create a feature branch `feature/$0-<short-slug>` from the default branch.
+3. Work the tasks in dependency order. Check each off in TASKS.md as you complete it.
+4. Write the tests each task's "Done when" calls for, alongside the change.
+5. Run the project's lint + typecheck + test commands locally.
+6. Open a PR whose body references epic `$0` and links the spec.
+
+## Rules
+
+- Implement only what the tasks call for. If code seems needed that no task covers, STOP and surface the gap — don't freelance scope.
+- Match existing code conventions; keep diffs small and reviewable.
+- No behavior change outside the spec's scope.
+- Every task ends done or explicitly deferred with a reason.
+
+## Output
+
+Code + tests on `feature/$0-<slug>`, PR opened. Write a short `docs/epics/$0/artifacts/IMPLEMENT-SUMMARY.md` listing completed tasks and the PR link so the AIDLC validator has a file to check.

package/dist/templates/speckit/skills/tasks.md

@@ -0,0 +1,38 @@
+---
+name: tasks
+description: Break a technical plan into an ordered, dependency-aware task list (TASKS.md), each task small, independently verifiable, and traceable to a requirement.
+argument-hint: "<{{EPIC_PREFIX}}-XXXX>"
+---
+
+# Tasks for Epic $0
+
+You are the **Tech Lead** agent.
+Load your full persona from `.claude/agents/aidlc-speckit-tech-lead.md` before starting.
+
+## Step 0: Pipeline Gate Check
+Read and execute `.claude/skills/_gate-check.md`. This skill = phase `tasks`, epic = `$0`. If gate fails → STOP.
+
+## Steps
+
+1. Read `docs/epics/$0/artifacts/PLAN.md` and `docs/epics/$0/artifacts/SPEC.md`.
+2. Decompose the plan into concrete, ordered tasks. Fill `docs/epics/$0/artifacts/TASKS.md`.
+
+## TASKS Contents
+
+- A numbered task list. For each task (`$0-T01`, …):
+  - **What**: the concrete unit of work.
+  - **Depends on**: task ids that must complete first (empty = can start immediately).
+  - **Parallel-safe**: yes/no — can run concurrently with its non-dependents.
+  - **Implements**: requirement id(s) / plan section it satisfies.
+  - **Done when**: the verifiable completion condition (incl. the test to write).
+
+## Rules
+
+- Right-size tasks — each independently verifiable, ideally one PR-sized change.
+- Order by dependency; mark what can proceed in parallel.
+- Every task references at least one requirement or plan section. No orphan tasks.
+- Every plan component must be covered by at least one task. No gaps.
+
+## Output
+
+Write the completed task list to `docs/epics/$0/artifacts/TASKS.md`.

package/dist/tools/epic-memory-hook.mjs

@@ -0,0 +1,78 @@
+#!/usr/bin/env node
+/**
+ * Claude Code UserPromptSubmit hook: when a prompt refers to an epic that has a
+ * memory digest, inject that digest into the context — so working on an epic
+ * automatically loads its prior decisions/constraints/reflections without you
+ * running /epic-context first.
+ *
+ * Opt-in: only active while enabled (aidlc / the extension writes the hook entry
+ * into ~/.claude/settings.json; disabling removes it). Zero-dep, and always
+ * exits 0 so it can never block a prompt.
+ *
+ * Contract: reads the hook payload JSON on stdin ({ prompt, cwd, ... }); on
+ * exit 0 whatever it prints to stdout is added to Claude's context.
+ */
+import fs from 'fs';
+import path from 'path';
+
+let raw = '';
+process.stdin.on('data', (c) => { raw += c; });
+process.stdin.on('end', () => {
+  try { run(raw); } catch { /* never break the prompt */ }
+  process.exit(0);
+});
+
+function run(raw) {
+  let input = {};
+  try { input = JSON.parse(raw) || {}; } catch { /* ignore */ }
+  const prompt = String(input.prompt || '');
+  const cwd = String(input.cwd || process.cwd());
+  if (!prompt.trim()) { return; }
+
+  const epicsDir = path.join(cwd, 'docs', 'epics');
+  let ids;
+  try {
+    ids = fs.readdirSync(epicsDir, { withFileTypes: true })
+      .filter((d) => d.isDirectory())
+      .map((d) => d.name);
+  } catch { return; } // no epics here
+  if (!ids.length) { return; }
+
+  const lower = prompt.toLowerCase();
+  const matched = ids.filter((id) => {
+    if (lower.includes(id.toLowerCase())) { return true; }
+    // Also match the leading code (DEMO-002 for DEMO-002-AUTO-REVIEW, EPIC-001 …).
+    const code = (id.match(/^[A-Za-z]+-\d+/) || [])[0];
+    return code ? lower.includes(code.toLowerCase()) : false;
+  });
+  if (!matched.length) { return; }
+
+  const blocks = [];
+  for (const id of matched) {
+    const p = path.join(epicsDir, id, 'epic-memory.json');
+    let mem;
+    try { mem = JSON.parse(fs.readFileSync(p, 'utf8')); } catch { continue; }
+    const digest = format(mem, id);
+    if (digest) { blocks.push(digest); }
+  }
+  if (!blocks.length) { return; }
+
+  console.log('# AIDLC epic memory (auto-loaded — prior context for this epic; prefer it over re-reading artifacts)\n');
+  console.log(blocks.join('\n\n'));
+}
+
+function format(m, id) {
+  const lines = [`## ${m.epic || id}`];
+  if (m.summary) { lines.push(`Summary: ${m.summary}`); }
+  const entries = Array.isArray(m.entries) ? m.entries : [];
+  const reflections = Array.isArray(m.reflections) ? m.reflections : [];
+  if (entries.length) {
+    lines.push('Decisions/constraints/context:');
+    for (const e of entries) { lines.push(`- [${e.kind || 'note'}] ${e.text}`); }
+  }
+  if (reflections.length) {
+    lines.push('Reflections (prompt/work better):');
+    for (const r of reflections) { lines.push(`- ${r.text}`); }
+  }
+  return lines.length > 1 ? lines.join('\n') : '';
+}