本文へ移動
cccskills
無料GitHub で公開

pmstudio-nfr

Detailed completeness audit of all project documents. Use when someone asks "are we ready for production", "audit our docs", "readiness check", "nfr tracker", or "what's missing". Do NOT use for a quick command listing — use /pmstudio instead. Scans every document for section-level completeness, reports staleness, and gives prioritized gap-closure recommendations.

インストール方法を見る

含まれるファイル(7)

  • SKILL.md5.8 KB
  • references/category-checklists.md8.6 KB
  • references/document-registry.md3.4 KB
  • templates/arb-submission.md4.3 KB
  • templates/dr-plan.md5.1 KB
  • templates/irp.md6.0 KB
  • templates/nfr-status-init.json579 B

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

NFR Tracker — Operational Readiness Auditor

Purpose

Meta-command that discovers all project documents, checks existence and freshness, and produces a readiness dashboard with prioritized recommendations for what to create or update next.

Process

Step 1: Detect Project Context

Read these files to understand the project:

  1. CLAUDE.local.md — project name, type, architecture, folder structure
  2. .sync-watch.json — registered documents and watch directories
  3. Run ls -R on the project root to discover all files

Determine the product lifecycle stage from context:

  • Ideate/Build — pre-production (DR/IRP not yet required)
  • Production — live product (DR/IRP required)
  • Mixed — some components live, some building (e.g., ProductA production + ProductB ideate)

Step 2: Scan Document Registry

Check for each document type. For each, record: exists (Y/N), path, last modified date, stale (Y/N based on threshold).

Document registry:

Document TypeSource SkillFile PatternRequired WhenStaleness
PRD/prd-generatorPRD/*.html, PRD/*.mdAlways30 days
Project Memory(auto)CLAUDE.local.mdAlways7 days
Presentation/project-docsPresentations/*.htmlRecommended30 days
Architecture Map/project-docsArchitecture/*.html, *Architecture*.htmlIf PRD has architecture section60 days
Stakeholder Directory/project-docsData/Stakeholder*.xlsx, Data/Stakeholder*.mdIf >5 stakeholders30 days
Change Log/change-logData/Change-Log*.mdRecommended14 days
ARB Presentation/arb-reviewPresentations/ARB-*.htmlBefore architecture review90 days
DR Plan/dr-planOperational/DR-Plan*.htmlProduction180 days
Incident Response Plan/irpOperational/IRP-*.htmlProduction180 days
Recovery Procedures/recovery-planOperational/Recovery-*.htmlIf DR Plan exists180 days
Verification Report/project-docsVerification/*.mdBefore milestones60 days

Staleness calculation: Compare file modification date to today. If (today - last_modified) > threshold, mark as STALE.

Search strategy: Use glob patterns first. If the project uses non-standard paths (from CLAUDE.local.md folder structure), adapt the search. Example: Some projects store PRDs in non-default paths like ProductB-Control-Framework/PRD/ not PRD/.

Step 3: Assess Completeness (Optional Deep Check)

If --deep flag or user requests "thorough check":

For each FOUND document, open it and check for key sections:

  • PRD: Has executive summary, requirements, timeline, risks? Missing sections = partial.
  • DR Plan: Has RTO/RPO targets, scenarios, procedures? Missing = partial.
  • IRP: Has severity matrix, escalation contacts, playbooks? Missing = partial.

Without --deep, existence + freshness is sufficient.

Step 4: Generate Recommendations

Prioritize missing/stale documents:

Priority logic:

  1. CRITICAL — Production product missing DR or IRP
  2. HIGH — Any required document missing or stale >2x threshold
  3. MEDIUM — Recommended document missing or stale >1x threshold
  4. LOW — Optional document missing

Each recommendation includes:

  • What to create/update
  • Why it matters
  • The slash command to run

Step 5: Output Report

Console output (default):

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
 NFR TRACKER ► READINESS REPORT — {Project Name}
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Lifecycle: {Production / Ideate / Mixed}
Scanned: {timestamp}

| Document              | Status  | Path              | Updated    | Stale? |
|-----------------------|---------|-------------------|------------|--------|
| ...                   | FOUND   | ...               | YYYY-MM-DD | No     |
| ...                   | MISSING | —                 | —          | —      |
| ...                   | FOUND   | ...               | YYYY-MM-DD | STALE  |

Coverage: N/M ({pct}%)    Stale: N    Missing: N

Recommendations:
1. [{priority}] {description}
   → Run: /{command}
...
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Markdown file output (with --format md):

Write to Operational/Readiness-Report-{YYYY-MM-DD}.md with the same content plus a header with project metadata.

Modes

  • /nfr-tracker — Full audit, console output
  • /nfr-tracker --format md — Write report to file
  • /nfr-tracker --deep — Check document completeness, not just existence
  • /nfr-tracker --scope {name} — Filter to a specific product/instance within the project

Critical Rules

  1. Read-only. This skill never creates or modifies documents. It only reports.
  2. Adapt paths. Don't assume standard folder structure. Read CLAUDE.local.md for actual paths.
  3. No false positives. Only mark FOUND if the file has real content (>100 bytes). Empty scaffolds don't count.
  4. Date accuracy. Use file system modification date, not content parsing, for freshness checks.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Train and optimize AI agents using Microsoft's Agent Lightning framework with reinforcement learning. Use when setting up agent training, instrumenting agents with tracing, configuring LightningStore, implementing reward functions, or optimizing prompts with RL/APO algorithms.

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5362026年10月11日 更新

Post-run self-evaluation system that scores agent output on correctness, clarity, actionability, and conciseness. Use after /team runs, skill executions, or when explicitly asked to evaluate output quality.

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5362026年10月11日 更新

Create AI marketing videos for ads, promos, product launches, and brand content. Models: Veo, Seedance, Wan, FLUX for visuals, Kokoro for voiceover. Types: product demos, testimonials, explainers, social ads, brand videos. Use for: Facebook ads, YouTube ads, product launches, brand awareness. Triggers: marketing video, ad video, promo video, commercial, brand video, product video, explainer video, ad creative, video ad, facebook ad video, youtube ad, instagram ad, tiktok ad, promotional video, launch video

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5362026年10月11日 更新

Use when building AI features into a product: LLM integration, RAG pipelines, guardrails, streaming, AI UX, prompt engineering, or AI cost control. Treats prompts as code and validates every model output.

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5362026年10月11日 更新

Your AI research and engineering brain trust. 59 named personas across 8 cells covering frontier labs, applied product, model architecture, reasoning/RL/agents, alignment and interpretability, theory and science of DL, multimodal and…

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5362026年10月11日 更新

Use when designing a new REST or GraphQL API, reviewing an API spec before implementation, setting team API standards, or migrating REST to GraphQL. Covers resources, HTTP semantics, pagination, error handling, and pitfalls.

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5362026年10月11日 更新

coco-research のスキルをすべて見る

このスキルの問題を報告する