qfai 1.0.3 → 1.0.5

package/assets/init/.qfai/assistant/instructions/constitution.md

@@ -0,0 +1,131 @@
+# QFAI Constitution (Non‑Negotiable)
+
+Updated: 2026-01-12
+
+This document defines **non‑negotiable operating rules** for QFAI agents and subagents.
+It is inspired by proven “constitution / articles / guardrails” patterns in existing SDD toolchains, but tailored to QFAI’s minimal workflow.
+
+---
+
+## Absolute Rule — Output Language
+
+**All outputs MUST be written in the user’s working language for this session.**
+
+- If the user writes in Japanese, output Japanese.
+- If the user writes in English, output English.
+- If the user mixes languages, prefer the dominant language unless explicitly instructed otherwise.
+
+This rule overrides all other stylistic preferences.
+
+---
+
+## Article I — Evidence over confidence
+
+Prefer **observable proof** over claims.
+When declaring completion, provide:
+
+- commands executed
+- key outputs (summaries; do not dump excessive logs)
+- exit codes / pass status
+
+If something cannot be verified in this environment, say so explicitly and proceed with the safest assumption.
+
+---
+
+## Article II — No invented facts
+
+Do **not** guess file paths, existing commands, or project policies.
+
+- Confirm with file search / grep / tree inspection before referencing.
+- If unknown, write `TBD` and record what evidence is missing.
+- If it blocks correctness, raise an Open Question.
+
+---
+
+## Article III — Project fit is mandatory (Project Memory)
+
+Before producing deliverables, read **project memory**:
+
+1. `.qfai/assistant/instructions/*`
+2. `.qfai/assistant/steering/*`
+3. `.qfai/require/require.md` (if present)
+4. `.qfai/specs/spec-*/` (if relevant)
+5. repository config (package.json, CI, scripts)
+
+Outputs MUST align with:
+
+- repository structure and conventions
+- chosen tools / runtimes
+- architecture boundaries
+
+---
+
+## Article IV — SDD is the source of truth
+
+If spec and code conflict:
+
+- fix the code to match the spec, OR
+- propose a spec change with rationale and accept it as a decision (do not silently drift)
+
+---
+
+## Article V — Traceability is mandatory
+
+Maintain traceability links:
+**Require → Spec → Scenario → Tests → Code → Verification evidence**
+
+Whenever practical, reference:
+
+- requirement IDs
+- spec section anchors
+- scenario titles
+
+---
+
+## Article VI — Clarification budget (avoid endless Q&A)
+
+Non-discussion commands MUST minimize questions.
+
+Default policy:
+
+- Ask **at most 5** clarifying questions total.
+- Prioritize **blocking** questions first.
+- If user requests `--auto`, proceed with explicit assumptions (label them).
+
+Stop conditions:
+
+- User says “stop / proceed / done”.
+- Question budget is exhausted.
+
+---
+
+## Article VII — Minimal scope with explicit deltas
+
+Make the smallest change that satisfies the spec and passes gates.
+If you must expand scope, declare it explicitly in a **Delta** section.
+
+---
+
+## Article VIII — Quality gates decide
+
+Do not claim “done” without passing the repo’s gate commands.
+Typical minimum (project-dependent):
+
+- format
+- lint
+- typecheck
+- tests
+- packaging verification (if distributed)
+
+---
+
+## Article IX — Preflight confidence gate (implementation/test stages)
+
+Before modifying code/tests, perform a **quick preflight**:
+
+- detect duplicate/overlapping implementations
+- confirm module boundaries and conventions
+- confirm where to update tests/docs
+- confirm how to run gates locally
+
+If confidence is low, ask targeted questions or run additional repo inspection.

package/assets/init/.qfai/assistant/instructions/workflow.md

@@ -0,0 +1,75 @@
+# QFAI Default Workflow
+
+Updated: 2026-01-12
+
+QFAI standardizes work into a fixed pipeline:
+
+**SDD → ATDD → TDD → Implementation → Verification**
+
+This file defines the canonical stages and delegation expectations.
+
+---
+
+## Absolute Rule — Output Language
+
+**All outputs MUST be written in the user’s working language for this session.**
+
+---
+
+## Stages (canonical)
+
+0. Steering refresh (project memory bootstrap)
+1. Discussion (optional): clarify idea → requirement seed
+2. Requirements: `.qfai/require/require.md`
+3. Specification (SDD): `.qfai/specs/spec-XXXX/`
+4. Scenario tests (ATDD): runnable scenario tests derived from `scenario.feature`
+5. Unit tests (TDD): runnable unit tests enforcing the spec
+6. Implementation: implement to satisfy spec + tests
+7. Verify: run quality gates and provide evidence
+
+---
+
+## Delegation pattern (multi‑role)
+
+A QFAI custom prompt may delegate to subagents (roles) and then consolidate results.
+
+Recommended delegation rules:
+
+- Delegate **analysis** and **review** (Architect / QA / Code Reviewer) early.
+- Delegate **contracts** only when needed (Contract Designer).
+- Delegate **CI/gates** verification to DevOps/CI Engineer when changes affect scripts or packaging.
+
+### Subagent response contract (required)
+
+When a subagent is invoked, they MUST respond using this structure:
+
+1. **Findings** (facts observed)
+2. **Recommendations** (what to do)
+3. **Proposed edits** (files/sections to change)
+4. **Open Questions / Risks**
+5. **Confidence** (High/Medium/Low + reason)
+
+---
+
+## Quality gates
+
+Gate commands are project-defined. Always discover them from the repo.
+Typical minimum:
+
+- format check
+- lint
+- typecheck
+- tests
+- pack/verify (if distributed)
+
+---
+
+## Evidence policy
+
+At the end of each stage, report:
+
+- what changed (file list)
+- what was executed (commands)
+- whether it passed (PASS/FAIL)
+
+Never claim completion without evidence.

package/assets/init/.qfai/assistant/prompts/README.md

@@ -0,0 +1,19 @@
+# prompts/
+
+SSOT prompt bodies used by tool-specific custom prompt definitions.
+
+Rule:
+
+- Tool-specific wrappers should be thin.
+- They should instruct the agent to read the corresponding file here and follow it.
+
+Files:
+
+- qfai-discuss.md (optional)
+- qfai-require.md
+- qfai-spec.md
+- qfai-scenario-test.md
+- qfai-unit-test.md
+- qfai-implement.md
+- qfai-verify.md
+- qfai-pr.md (optional)

package/assets/init/.qfai/assistant/prompts/qfai-discuss.md

@@ -0,0 +1,173 @@
+<!--
+QFAI Prompt Body (SSOT)
+- This file is intended to be referenced by tool-specific custom prompt definitions (e.g., Copilot .prompt.md, Claude Code slash commands).
+- Keep tool-specific wrappers thin: "Read this file and follow it."
+-->
+
+---
+
+id: qfai-discuss
+title: QFAI Discuss (Idea → Clear Requirements)
+description: "Socratic discussion to turn a vague idea into a clear, testable set of requirements inputs."
+argument-hint: "<idea-or-problem> [--auto]"
+allowed-tools: [Read, Glob, Write, TodoWrite, Task, Bash]
+roles: [Facilitator, Interviewer, RequirementsAnalyst, QAEngineer, Planner]
+mode: interactive-by-default
+
+---
+
+# /qfai-discuss — Discussion → Requirements Clarity
+
+## Purpose
+
+Use this when the user has only an idea in their head. Your job is to **make the requirements explicit and testable** with minimal user burden.
+
+## Success Criteria (Definition of Done)
+
+- A “Requirements Seed” exists: goals, non-goals, constraints, acceptance criteria (high level), and open questions.
+- The output is ready to be fed into **/qfai-require** with minimal further clarification.
+
+## Non‑Negotiable Principles (QFAI Articles)
+
+These principles are inspired by “constitution / articles” patterns used by other agent frameworks, but tailored to QFAI.
+
+1. **SDD First (Specification is the source of truth)**  
+   If there is a conflict between code and spec, treat the spec as authoritative and either (a) fix code or (b) raise an explicit Open Question to change the spec.
+
+2. **Traceability is mandatory**  
+   Every meaningful change must be traceable: **Require → Spec → Scenario → Tests → Code → Verification evidence**.
+
+3. **Evidence over confidence**  
+   Prefer observable proof (logs, commands, file diffs, test results). If you cannot verify, say so and record it.
+
+4. **Minimize scope, but never hide gaps**  
+   Keep changes minimal, but do not “paper over” missing decisions. If something blocks correctness, stop and ask.
+
+5. **Quality gates are the decision mechanism**  
+   Use tests/lint/typecheck/build/pack verification (whatever the repo defines) as the primary guardrail. Fix until PASS.
+
+6. **Make it runnable**  
+   Outputs must be executable in terminal/CI. Provide copy‑paste commands.
+
+7. **User time is expensive**  
+   Ask only the questions that are truly blocking. Everything else: make reasonable assumptions and label them clearly.
+
+## Absolute Rule — Output Language
+
+**All outputs MUST be written in the user’s working language for this session.**
+
+- If the user writes in Japanese, output Japanese.
+- If the user writes in English, output English.
+- If the user mixes languages, prefer the dominant language unless explicitly instructed otherwise.  
+  This rule overrides all other stylistic preferences.
+
+## Multi‑Role Orchestration (Subagents)
+
+This workflow assumes the environment _may_ support subagents (e.g., Claude Code “Task” tool) or may not.
+
+### If subagents are supported
+
+Delegate to multiple roles and then merge the results. Use a “real‑world workflow” order:
+
+- Facilitator → Interviewer → Requirements Analyst → Planner → Architect → (Contract Designer) → Test Engineer → QA Engineer → Code Reviewer → DevOps/CI Engineer
+
+**Pseudo‑invocation pattern** (adjust to your tool):
+
+```text
+Task(
+  subagent_type="planner",
+  description="Create an execution plan and DoD",
+  prompt="Context: ...\nGoal: ...\nConstraints: ...\nReturn: phases + risks + DoD"
+)
+```
+
+### If subagents are NOT supported
+
+Simulate roles by running the same sequence yourself:
+
+- Write a short “role output” section per role, then consolidate into the final deliverable(s).
+
+## Step 0 — Load Context (always)
+
+1. Read relevant **project steering** (if present):
+   - `.qfai/assistant/steering/structure.md`
+   - `.qfai/assistant/steering/tech.md`
+   - `.qfai/assistant/steering/product.md`
+   - any additional files under `.qfai/assistant/steering/`
+
+2. Read **project constitution / instructions** (if present):
+   - `.qfai/assistant/instructions/constitution.md`
+   - `.qfai/assistant/instructions/workflow.md` (or equivalent)
+
+3. Read existing artifacts for the current work item (if present):
+   - `.qfai/require/`
+   - `.qfai/specs/spec-*/`
+   - `.qfai/contracts/`
+
+4. Inspect repo conventions:
+   - package manager (pnpm/npm/yarn), test runner, lint/typecheck scripts, CI definitions
+   - existing test patterns (unit/integration/e2e)
+
+## Step 1 — Frame the discussion (Facilitator)
+
+Produce a short framing first (no more than ~10 lines):
+
+- Problem statement
+- Target users / stakeholders
+- Expected outcome
+- Scope boundary (in / out)
+- Constraints (time, platform, compatibility posture)
+
+## Step 2 — Ask only high‑value questions (Interviewer)
+
+Generate questions in **priority order**:
+
+- **Blockers**: must be answered to write requirements
+- **Clarifiers**: improve precision but can be assumed temporarily
+
+Use a _Socratic style_:
+
+- Ask one question at a time in interactive mode.
+- If `--auto` is provided, make explicit assumptions and mark them.
+
+## Step 3 — Draft the Requirements Seed (Requirements Analyst)
+
+Write a draft in this format:
+
+### Requirements Seed
+
+- **Goal**:
+- **Non‑Goals**:
+- **Users / Actors**:
+- **Key User Journeys** (1–3):
+- **Constraints**:
+- **Acceptance Criteria (high level)**:
+- **Observability** (what evidence proves success):
+- **Risks / Edge cases**:
+- **Assumptions**:
+- **Open Questions (blockers)**:
+- **Open Questions (non‑blockers)**:
+
+## Step 4 — QA sanity check (QA Engineer)
+
+Validate:
+
+- Acceptance criteria are testable.
+- Failure modes are considered.
+- Observability is defined (logs/messages/output).
+
+## Step 5 — Produce handoff to /qfai-require (Planner)
+
+Generate the minimal input payload for /qfai-require:
+
+- Short summary
+- Confirmed facts
+- Remaining questions (if any)
+- Proposed requirement ID namespace (optional)
+
+## Output
+
+Return:
+
+1. Requirements Seed (as above)
+2. The “/qfai-require input” block (copy‑paste ready)

package/assets/init/.qfai/assistant/prompts/qfai-implement.md

@@ -0,0 +1,239 @@
+<!--
+QFAI Prompt Body (SSOT)
+- This file is intended to be referenced by tool-specific custom prompt definitions (e.g., Copilot .prompt.md, Claude Code slash commands).
+- Keep tool-specific wrappers thin: "Read this file and follow it."
+-->
+
+---
+
+id: qfai-implement
+title: QFAI Implement (Spec-driven implementation)
+description: "Implement the program feature according to specs/contracts/scenario; includes tests, review, and full quality gate."
+argument-hint: "<spec-id> [--auto]"
+allowed-tools: [Read, Glob, Write, TodoWrite, Task, Bash]
+roles: [Planner, Architect, BackendEngineer, FrontendEngineer, TestEngineer, QAEngineer, CodeReviewer, DevOpsCIEngineer]
+mode: iterative
+
+---
+
+# /qfai-implement — Implement Feature (Spec‑Driven)
+
+## Purpose
+
+Implement the required feature/changes according to **spec + contracts + scenario**, then reach a **green quality gate**.
+
+## Success Criteria (Definition of Done)
+
+- Implementation matches the spec and contracts.
+- Scenario tests + unit tests pass.
+- Repo quality gates pass (lint/type/build/pack as applicable).
+- Verification evidence is recorded (commands + results).
+
+## Non‑Negotiable Principles (QFAI Articles)
+
+These principles are inspired by “constitution / articles” patterns used by other agent frameworks, but tailored to QFAI.
+
+1. **SDD First (Specification is the source of truth)**  
+   If there is a conflict between code and spec, treat the spec as authoritative and either (a) fix code or (b) raise an explicit Open Question to change the spec.
+
+2. **Traceability is mandatory**  
+   Every meaningful change must be traceable: **Require → Spec → Scenario → Tests → Code → Verification evidence**.
+
+3. **Evidence over confidence**  
+   Prefer observable proof (logs, commands, file diffs, test results). If you cannot verify, say so and record it.
+
+4. **Minimize scope, but never hide gaps**  
+   Keep changes minimal, but do not “paper over” missing decisions. If something blocks correctness, stop and ask.
+
+5. **Quality gates are the decision mechanism**  
+   Use tests/lint/typecheck/build/pack verification (whatever the repo defines) as the primary guardrail. Fix until PASS.
+
+6. **Make it runnable**  
+   Outputs must be executable in terminal/CI. Provide copy‑paste commands.
+
+7. **User time is expensive**  
+   Ask only the questions that are truly blocking. Everything else: make reasonable assumptions and label them clearly.
+
+## Absolute Rule — Output Language
+
+**All outputs MUST be written in the user’s working language for this session.**
+
+- If the user writes in Japanese, output Japanese.
+- If the user writes in English, output English.
+- If the user mixes languages, prefer the dominant language unless explicitly instructed otherwise.  
+  This rule overrides all other stylistic preferences.
+
+## Multi‑Role Orchestration (Subagents)
+
+This workflow assumes the environment _may_ support subagents (e.g., Claude Code “Task” tool) or may not.
+
+### If subagents are supported
+
+Delegate to multiple roles and then merge the results. Use a “real‑world workflow” order:
+
+- Facilitator → Interviewer → Requirements Analyst → Planner → Architect → (Contract Designer) → Test Engineer → QA Engineer → Code Reviewer → DevOps/CI Engineer
+
+**Pseudo‑invocation pattern** (adjust to your tool):
+
+```text
+Task(
+  subagent_type="planner",
+  description="Create an execution plan and DoD",
+  prompt="Context: ...\nGoal: ...\nConstraints: ...\nReturn: phases + risks + DoD"
+)
+```
+
+### If subagents are NOT supported
+
+Simulate roles by running the same sequence yourself:
+
+- Write a short “role output” section per role, then consolidate into the final deliverable(s).
+
+## Behavior Rules (high leverage)
+
+- **Language**: Output MUST follow the user’s working language for this session.
+- **Question budget**: Ask at most **5** clarifying questions total. Prioritize blockers. If `--auto`, proceed with explicit assumptions.
+- **No hallucination**: Do not invent file paths, commands, or policies. Confirm via repo inspection.
+- **Evidence**: Do not claim completion without commands/results (format/lint/type/test/pack as applicable).
+- **Subagent contract**: When delegating, require the subagent response structure:
+  1. Findings 2) Recommendations 3) Proposed edits 4) Open Questions/Risks 5) Confidence
+
+## Step 0 — Load Context (always)
+
+1. Read relevant **project steering** (if present):
+   - `.qfai/assistant/steering/structure.md`
+   - `.qfai/assistant/steering/tech.md`
+   - `.qfai/assistant/steering/product.md`
+   - any additional files under `.qfai/assistant/steering/`
+
+2. Read **project constitution / instructions** (if present):
+   - `.qfai/assistant/instructions/constitution.md`
+   - `.qfai/assistant/instructions/workflow.md` (or equivalent)
+
+3. Read existing artifacts for the current work item (if present):
+   - `.qfai/require/`
+   - `.qfai/specs/spec-*/`
+   - `.qfai/contracts/`
+
+4. Inspect repo conventions:
+   - package manager (pnpm/npm/yarn), test runner, lint/typecheck scripts, CI definitions
+   - existing test patterns (unit/integration/e2e)
+
+## Step 0 — Project Analysis (mandatory)
+
+Before producing any deliverable, **thoroughly analyze the current project** so your outputs fit the repo’s:
+
+- background and goals
+- directory structure and conventions
+- chosen technologies and versions (runtime, package manager, test runner)
+- architecture boundaries (packages, CLI, core modules)
+- existing patterns for tests, docs, and CI
+
+### Minimum analysis checklist
+
+- [ ] Read key repo docs: README / CHANGELOG / RELEASE (if present)
+- [ ] Inspect `.qfai/` layout and existing SDD/ATDD/TDD artifacts (if present)
+- [ ] Inspect `packages/qfai` structure (CLI entrypoints, core modules, validators, assets/init)
+- [ ] Identify standard gate commands (format/lint/type/test/verify-pack) and where they are defined
+- [ ] Search for existing examples/patterns of similar changes in tests (if available)
+- [ ] Note constraints: Node versions, CI matrix, packaging rules, verify-pack expectations
+
+If analysis cannot be performed (missing access), clearly state what could not be verified and proceed with minimal-risk assumptions.
+
+## Step 0.5 — Steering Bootstrap / Refresh (mandatory when incomplete)
+
+QFAI expects `assistant/steering/` to contain **project‑specific facts** so all subsequent design/test/implementation fits this repository.
+
+### What to do
+
+1. Open these files:
+
+- `.qfai/assistant/steering/product.md`
+- `.qfai/assistant/steering/tech.md`
+- `.qfai/assistant/steering/structure.md`
+
+2. If they are missing, mostly empty, or still have placeholders (e.g., `- ` only), **populate them by analyzing the current repository**:
+
+- derive “what/why/users/success/non-goals” from README/docs/issues (product.md)
+- derive runtime/tooling versions + constraints from package.json, CI config, lockfiles (tech.md)
+- derive repo layout + key directories + gate commands from the file tree and scripts (structure.md)
+
+3. Do **not** invent facts. If something cannot be verified, write it as:
+
+- `TBD` + what evidence is missing, or
+- an Open Question (if it blocks correctness)
+
+### Steering refresh checklist
+
+- [ ] product.md: what we build / users / success / non-goals / release posture
+- [ ] tech.md: Node / package manager / TS / test / lint / CI constraints
+- [ ] structure.md: repo layout, key packages, entrypoints, standard gate commands, how to run locally
+
+## Step 1 — Confirm prerequisites
+
+Must exist:
+
+- `.qfai/specs/spec-XXXX/spec.md`
+- `.qfai/specs/spec-XXXX/delta.md`
+- `.qfai/specs/spec-XXXX/scenario.feature`
+  If missing, stop and request /qfai-spec.
+
+## Step 2 — Plan the implementation (Planner + Architect)
+
+Create a short plan:
+
+- Tasks (ordered)
+- Files likely affected
+- Risks + mitigations
+- Definition of Done (commands that must pass)
+
+If tool supports TodoWrite, record tasks.
+
+## Step 3 — Implement in small increments (Engineers)
+
+Rules:
+
+- Prefer small, reviewable commits (even if local).
+- Keep changes minimal and aligned with spec.
+- If spec is ambiguous, do not guess silently: record an Open Question and/or propose a spec update.
+
+## Step 4 — Keep tests in lockstep (Test Engineer)
+
+- If tests exist, update them only when spec changes.
+- If tests are missing, add the minimal tests needed to enforce the spec.
+
+## Step 5 — Review & QA checks
+
+- Code Reviewer reviews diffs for maintainability and risk.
+- QA Engineer checks acceptance criteria coverage and failure handling.
+
+## Step 6 — Run quality gates (DevOps/CI Engineer)
+
+Run the repo’s standard commands. At minimum:
+
+- formatting
+- lint
+- typecheck (if applicable)
+- unit tests
+- scenario tests (if applicable)
+
+Record:
+
+- commands
+- outputs (summary)
+- PASS/FAIL
+
+## Step 7 — If any gate fails: fix loop
+
+Iterate until all gates pass, prioritizing:
+
+1. correctness vs spec
+2. test determinism
+3. maintainability
+
+## Output
+
+- Implementation diffs
+- Updated tests (if needed)
+- Verification evidence (commands + results)
+- Suggested next command: /qfai-verify (if not already done)