bmad-readiness-check — скилл для Claude Code: проверка и установка
Скилл bmad-readiness-check для Claude Code сверяет документы планирования перед началом разработки. Запускается по просьбе проверить готовность требований, архитектуры и эпиков к реализации.
MIT 107 установок ★ 493 обновлён 2026-06-20
npx skills add aj-geddes/claude-code-bmad-skills --skill bmad-readiness-check -a claude-codeЧто делает bmad-readiness-check
Скилл проверяет, согласованы ли требования, архитектура и эпики или истории. Он ищет пробелы в покрытии функциональных и нефункциональных требований, проверяет связи эпиков с требованиями и отмечает недочёты архитектуры. На выходе — вердикт PASS, CONCERNS или FAIL с конкретными замечаниями.
Сначала скрипт ищет файлы планирования в bmad-output/ или в заданной папке. Затем скилл сопоставляет документы и сохраняет отчёт readiness-report-<project-slug>-<date>.md. Для этого нужны документы PRD или tech-spec и архитектуры; эпики и истории проверяются, если они есть. Скилл работает только с документами: код не меняет, тесты и сборку не запускает.
Какие задачи решает
- сверить функциональные требования с архитектурой
- найти требования без архитектурного покрытия
- проверить связь эпиков с требованиями
- выявить пропуски в документах перед разработкой
Примеры запросов агенту
- Проверь, готовы ли мы начинать разработку.
- Сверь PRD с архитектурой и найди пробелы.
- Проведи проверку готовности планирования.
Как установить bmad-readiness-check
1Через Skills CLI (Vercel)
npx skills add aj-geddes/claude-code-bmad-skills --skill bmad-readiness-check -a claude-code2Вручную: положить файлы
~/.claude/skills/bmad-readiness-check/SKILL.mdПапка скилла целиком; для одного проекта — .claude/skills/bmad-readiness-check/
1Через Skills CLI (Vercel)
npx skills add aj-geddes/claude-code-bmad-skills --skill bmad-readiness-check -a codex2Вручную: положить файлы
.agents/skills/bmad-readiness-check/SKILL.mdДля всех проектов — ~/.agents/skills/
1Через Skills CLI (Vercel)
npx skills add aj-geddes/claude-code-bmad-skills --skill bmad-readiness-check -a cursor2Вручную: положить файлы
.cursor/skills/bmad-readiness-check/SKILL.mdCursor читает и .agents/skills/, и .claude/skills/
1GitHub CLI
gh skill install aj-geddes/claude-code-bmad-skills bmad-planning-orchestrator/skills/bmad-readiness-check2Вручную: положить файлы
.github/skills/bmad-readiness-check/SKILL.md1Через Skills CLI (Vercel)
npx skills add aj-geddes/claude-code-bmad-skills --skill bmad-readiness-check -a gemini-cli2Командой платформы
gemini skills install https://github.com/aj-geddes/claude-code-bmad-skills --path bmad-planning-orchestrator/skills/bmad-readiness-checkВручную: положить файлы
.devin/skills/bmad-readiness-check/SKILL.mdСтарый путь .windsurf/skills/ тоже читается
Через Skills CLI (Vercel)
npx skills add aj-geddes/claude-code-bmad-skills --skill bmad-readiness-check -a opencodeClaude.ai
Упакуйте папку в zip и загрузите в Settings → Capabilities → Skills
По шагам, с командами для каждой платформы: Как установить скилл в Claude Code: пять способов
Вопросы о bmad-readiness-check
Как установить bmad-readiness-check в Claude Code?
Выполни npx skills add aj-geddes/claude-code-bmad-skills --skill bmad-readiness-check -a claude-code. Файл скилла устанавливается по пути ~/.claude/skills/bmad-readiness-check/SKILL.md.
Что делает bmad-readiness-check?
Скилл сопоставляет требования, архитектуру и эпики или истории, а затем выдаёт вердикт PASS, CONCERNS или FAIL и сохраняет отчёт о готовности.
Какие документы проверяет bmad-readiness-check?
Он проверяет PRD или tech-spec, архитектурный документ, а при наличии — эпики и истории. Для Enterprise также могут учитываться дополнения по безопасности и DevOps.
Исходный текст
SKILL.md · MIT
---
name: bmad-readiness-check
description: |
SOLUTIONING GATE — validates cohesion across planning artifacts before any
implementation begins. Cross-references PRD (or tech-spec), architecture document,
and epics/stories for coverage consistency and missing pieces. Returns a verdict
of PASS, CONCERNS, or FAIL with specifics. Does NOT touch code.
Use when the user says:
- "check if we're ready to build" / "are we ready to implement?"
- "validate planning" / "gate check" / "readiness check"
- "run the solutioning gate" / "check implementation readiness"
- "is the architecture complete?" / "do requirements match architecture?"
- "sign off on planning" / "planning done, check it"
- "validate PRD against architecture" / "cross-reference planning docs"
Intents supported: Validate (primary).
Track-adaptive: works for Quick Flow (tech-spec + architecture), BMad Method
(PRD + architecture + epics), and Enterprise (all of the above + optional
security/DevOps addenda).
allowed-tools: Read, Write, Bash, Glob, Grep, TodoWrite
---
# BMAD Readiness Check
**"Planning ends here."** This skill is the gate between Solutioning and
Implementation. It validates that the planning corpus is internally consistent
— requirements are covered by architecture, epics trace back to requirements,
and nothing critical is missing — before a single line of code is written.
The output is a `readiness-report.md` with a clear PASS / CONCERNS / FAIL
verdict and actionable specifics.
---
## Workflow
Use TodoWrite to track: Load Artifacts → Cross-Reference → Quality Checks →
Generate Report → Display Verdict.
---
### Step 1 — Locate Planning Artifacts
Run the bundled existence check:
```bash
bash "${CLAUDE_PLUGIN_ROOT}/skills/bmad-readiness-check/scripts/readiness-check.sh" <output-folder>
```
The script checks for required artifact files and prints a PASS / CONCERNS /
FAIL pre-flight verdict. Read its output; it will also print artifact paths
for you to load in Step 2.
Default search root is `bmad-output/` (or the user-configured `outputFolder`).
If the user supplies a custom path, pass it as the argument.
---
### Step 2 — Load and Parse Artifacts
Read each artifact the script located:
**Requirements document** (PRD or tech-spec):
- Extract labelled Functional Requirements (FR-001, FR-002 …) or requirement
bullets if not labelled.
- Extract Non-Functional Requirements (NFR-001 … or named sections:
Performance, Security, Scalability, Reliability, Maintainability).
- Note epic count and high-level scope statements.
**Architecture document**:
- Extract system components and their responsibilities.
- Locate FR traceability matrix or explicit FR-to-component mappings.
- Locate NFR coverage sections.
- Note technology stack decisions and trade-off notes.
**Epics / stories** (if present):
- Count epics and stories.
- For each epic, identify which PRD requirement(s) it references.
Record a baseline:
```
Baseline
- FRs found: N
- NFRs found: N
- Epics found: N (0 if Quick Flow)
- Stories found: N (0 if not yet decomposed)
```
---
### Step 3 — Cross-Reference Checks
#### 3a. FR → Architecture Coverage
For each FR, search the architecture document for the FR identifier AND for
the subject matter of the requirement. Mark:
- **Covered** — FR explicitly addressed, component assigned.
- **Implied** — subject matter covered but not explicitly linked.
- **Missing** — no evidence of coverage.
#### 3b. NFR → Architecture Coverage
For each NFR category, check for a dedicated architecture section or explicit
strategy:
- Performance: caching, response-time targets, async patterns.
- Security: auth/authz model, encryption, secrets management.
- Scalability: horizontal/vertical strategy, load balancing.
- Reliability: failover, backup, uptime targets.
- Others as present.
Mark each: **Addressed** / **Partial** / **Missing**.
#### 3c. Epic → FR Traceability (BMad Method / Enterprise tracks)
For each epic, verify it references at least one FR or PRD section.
Flag orphan epics (no traceable requirement).
#### 3d. Architecture Quality Spot-Checks
Verify the architecture document contains:
- [ ] Architectural pattern stated and justified
- [ ] Component responsibilities and interfaces defined
- [ ] Data model / entity relationships
- [ ] API design or service contracts
- [ ] Technology choices with rationale
- [ ] Trade-offs documented
- [ ] Assumptions and constraints listed
---
### Step 4 — Compute Verdict
Apply these thresholds (adapted from BMAD Method v6 gate criteria, expressed
as PASS / CONCERNS / FAIL):
| Criterion | PASS | CONCERNS | FAIL |
|-----------|------|----------|------|
| FR coverage (covered + implied) | ≥ 90 % | 80–89 % | < 80 % |
| NFR coverage (addressed + partial) | ≥ 90 % | 80–89 % | < 80 % |
| Architecture quality checks | ≥ 80 % | 70–79 % | < 70 % |
| Blocker issues (critical gaps) | 0 | have mitigation | unmitigated |
| Epic traceability (if applicable) | all linked | ≥ 80 % linked | < 80 % |
**Overall verdict rule:**
- **PASS** — all criteria at PASS threshold, no blockers.
- **CONCERNS** — one or more criteria in CONCERNS band, blockers have
mitigation plans. Proceed with caution; address concerns during story
refinement.
- **FAIL** — any criterion below CONCERNS threshold, or any unmitigated
blocker. Do not proceed to story creation until issues are resolved.
---
### Step 5 — Generate Report
Write the readiness report using the template:
```
${CLAUDE_PLUGIN_ROOT}/skills/bmad-readiness-check/templates/readiness-report.template.md
```
Save to: `<outputFolder>/readiness-report-<project-slug>-<date>.md`
Use the Write tool. Fill all `{{placeholder}}` fields from your analysis.
---
### Step 6 — Display Verdict
Print a concise summary to the user:
```
Readiness Check Complete
Artifact Coverage
FR coverage: XX %
NFR coverage: XX %
Architecture quality: XX %
Verdict: PASS | CONCERNS | FAIL
<One-sentence rationale>
Report saved: <file path>
```
Then give a concrete **Next Step** recommendation:
- **PASS** → Proceed to epic/story decomposition
(`/bmad-planning-orchestrator:bmad-epics-and-stories`).
- **CONCERNS** → Proceed with caution; list the open items the story author
must carry forward as Dev Notes.
- **FAIL** → List the top 3–5 required fixes; re-run readiness check after
addressing them.
---
## Scope Boundary
This skill validates **planning documents only**. It does not:
- Write application code.
- Run tests, linters, or build tools.
- Review pull requests or diffs.
- Execute coverage or velocity metrics.
When the verdict is PASS or CONCERNS, hand off to your dev tools via the
story files produced by the epic/story skill.
---
## Persona Note
This gate is traditionally owned by Winston (System Architect) in the BMAD
Method. The skill is a workflow, not a character, but Winston's voice —
thorough, systematic, quality-focused — should guide the tone of the report.
---
> ---
> Part of the **BMAD Planning & Orchestrator** plugin — a Claude Code harness for the **BMAD Method** by the **BMAD Code Organization** (https://github.com/bmad-code-org/BMAD-METHOD). Implements the spirit of `bmad-check-implementation-readiness`. All methodology credit belongs to the BMAD Code Organization.
© aj-geddes · MIT · aj-geddes/claude-code-bmad-skills