mcpField
D
docs-check

docs-check

github.com/oliver-kriskaTesting & quality57145 installsUpdated 1w ago
Repository

About docs-check

docs-check is a Claude Code skill for checking plugin agents, skills, hooks, and configuration against current documentation. It combines cached documentation comparisons with local structural checks and can limit validation by component.

Key features

  1. 01Validate agents, skills, hooks, and configuration
  2. 02Fetch and cache Claude Code documentation
  3. 03Quick mode for structural checks only
  4. 04Focus checks on selected components
  5. 05Report issues and offer a PR

Use cases

  1. 01Check a plugin for breaking documentation changes
  2. 02Run local structural checks without fetching docs
  3. 03Review only hooks or configuration
  4. 04Find new features applicable to plugin components

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/oliver-kriska/claude-elixir-phoenix --skill docs-check

After installing, type/docs-checkin Claude Code to confirm it loaded.

Bundled filesreferences/doc-pages.mdreferences/validation-rules.md

Details

From the project README

Plugin Documentation Compatibility Check

Validates plugin agents, skills, hooks, and config against the latest Claude Code documentation to catch breaking changes and discover new features.

Usage

/docs-check                    # Full validation (all components)
/docs-check --quick            # Structural checks only (no docs fetch, no tokens)
/docs-check --focus=agents     # Validate only agents
/docs-check --focus=skills     # Validate only skills
/docs-check --focus=hooks      # Validate only hooks
/docs-check --focus=config     # Validate only plugin.json/marketplace.json

Architecture (OTP Supervision Pattern)

┌─────────────────────────────────────────────────────────────────┐
│  /docs-check (skill entry point)                                │
│   │                                                             │
│   ├─ Step 1: bash scripts/fetch-claude-docs.sh (zero tokens)    │
│   │          Always fetches all 9 doc pages (~420KB)             │
│   │                                                             │
│   └─ Step 2: delegate to orchestrator (reads from cache only)   │
│       │                                                         │
│       │  docs-validation-orchestrator (opus)                    │
│       │                                                         │
│       │  SCAN → READ CACHE → SPAWN WORKERS → COMPRESS → REPORT │
│       │   │         │              │             │          │   │
│       │   ↓         ↓              ↓             ↓          ↓   │
│       │ inventory  pre-fetched  4 parallel    context    report │
│       │ plugin     docs-cache   subagents     supervisor       │
│       │ components              (sonnet)      (haiku)          │
│       └─────────────────────────────────────────────────────────┘
└─────────────────────────────────────────────────────────────────┘

Execution

Step 1: Fetch Docs (Automatic)

Always run first. Downloads all doc pages to cache. Skips pages already cached within 24h. Zero token cost — pure curl.

# --quick mode: skip this step entirely (structural checks only)
# All other modes: always fetch
bash scripts/fetch-claude-docs.sh
Step 2: Delegate to Orchestrator

After docs are cached, delegate. The orchestrator reads from cache only and crashes if cache files are missing.

Task(subagent_type: "docs-validation-orchestrator")

Pass the user's flags (--quick, --focus) in the prompt.

What the Orchestrator Does

  1. Inventory — scan plugins/elixir-phoenix/ for existing components
  2. Read cached docs — from .claude/docs-check/docs-cache/ (never fetches)
  3. Spawn workers — one sonnet subagent per component type, in parallel
  4. Compress — context-supervisor (haiku) if 3+ workers
  5. Structural checks — fast local checks, always run
  6. Report & Action — write report, offer PR if issues found

Iron Laws

  1. Fetch ALL docs upfront — no conditional fetching, no partial downloads
  2. Use scripts/fetch-claude-docs.sh — single source of truth for doc fetching
  3. Workers get docs IN PROMPT — no runtime fetching
  4. Workers use sonnet — opus is wasteful for comparison tasks
  5. Structural checks always run — even if docs fetch fails
  6. Breaking changes are BLOCKERS — surface prominently

References

  • references/validation-rules.md — Per-component validation checklists
  • references/doc-pages.md — Component-to-URL mapping

Comments

Markdown supported