By stgmt
Team coding standards and workflows for Claude Code: skills (create-spec, run-tests, plan-pomogator, etc.), rules, hooks, and MCP servers — distributed via canonical Anthropic marketplace.
Настроить whitelist файлов в корне репозитория
Диагностика окружения dev-pomogator: GitHub CLI, Node/Git/Bun/Python/MCP/hooks/env vars с предложением переустановки
Просмотр и управление очередью автозахваченных сигналов
Prepare a sanitized GitHub issue draft and submit it only after approval
Анализ сессии и предложение Claude rules на основе выявленных паттернов
Keep spec/markdown cross-reference links clickable in the editor (Marksman) when you RENAME a heading. Long descriptive headings (`## FR-7: Title`) are Marksman's standard, but their GLFM slug is derived from the heading TEXT — so renaming a heading silently breaks every inbound `[text](file.md#old-slug)` link. This skill DETECTS broken anchors and AUTO-FIXES them: ~99% deterministically (id-bearing links), the rest via `claude -p`. INVOKE after renaming/retitling a heading, before declaring spec work done, or to clean a corpus of broken anchors. Triggers (RU): "почини якоря", "битые ссылки в спеках", "переименовал заголовок ссылки", "anchor fix", "проверь ссылки спек". Triggers (EN): "fix broken anchors", "spec links broken after rename", "check spec link integrity", "anchor integrity".
Аудит черновика на жаргон, внутренние коды, multi-select >3 опций и отсутствие микроистории. Триггеры (RU): "проверь черновик", "ревью ответа", "answer simple"; (EN): "audit draft", "check response", "answer-simple draft". Slash invocation `/answer-simple <черновик>` возвращает блоки "Переформулировано:" и "Найдено проблем:". Без аргумента — usage hint.
Autonomous fix→verify→re-verify loop for architecture-decision-builder. Runs the whole verification battery via one driver (arch-review.ts), triages findings, fixes the mechanical ones, re-runs, and repeats until PASS — WITHOUT a human prompting each cycle. Triggers: "прогони цикл починок", "self-review loop", "verify-fix loop", "погоняй проверки сам", after editing any architecture-decision helper/spec, or before declaring the skill done.
Permanently configure Microsoft Edge on Windows to launch with --remote-debugging-port=9222 across taskbar, Start Menu, Quick Launch shortcuts and MSEdgeHTM/MSEdgeMHT/MSEdgePDF/microsoft-edge: registry handlers. Lets Playwright/Puppeteer chromium.connectOverCDP attach to the user's REAL Edge profile (cookies, extensions, sessions, open tabs) instead of spawning an empty disposable Edge instance via chromium.launch({channel:'msedge'}). Use this skill BEFORE writing browser automation that needs the user's authenticated state, OR when an agent has been told not to open empty Edge windows. Reversible via -Revert.
HOW and WHY to make a FAST model call (Haiku/Sonnet) from our own hooks, gates, judges, and engine scripts — via the local Meridian subscription proxy on http://127.0.0.1:3456, NOT via `claude -p`. Meridian gives an Anthropic-compatible /v1/messages API backed by the user's SUBSCRIPTION (no ANTHROPIC_API_KEY needed). Measured: a direct Meridian call with thinking OFF ≈ 2.4s; `claude -p` ≈ 13s (it cold-starts MCP/hooks/plugin every call). USE THIS before writing any "spawn claude to judge X" code — do not reinvent the slow wheel. Triggers: "вызвать модель из хука", "быстрый вызов хайку", "судья на хайку", "model call from a hook/gate/judge", "call haiku fast", "LLM judge in a hook", "meridian api call", "как звать модель в плагине".
Matches all tools
Hooks run on every tool call, not just specific ones
Admin access level
Server config contains admin-level keywords
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
This plugin requires configuration values that are prompted when the plugin is enabled. Sensitive values are stored in your system keychain.
spec_access_enforceWhen ON (default), the agent's raw Read/Grep/Edit/Write/Bash over .specs/** is DENIED and routed through the dev-pomogator-specs MCP door (read_spec_doc / apply_spec_change / etc.); deliberate bypass via SPEC_ACCESS_SKIP=1. Turn OFF only to downgrade to shadow (log-only). Lets you get enforcement without hand-editing settings.json.
${user_config.spec_access_enforce}Executes bash commands
Hook triggers when Bash tool is used
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Modifies files
Hook triggers on file write and edit operations
Canonical Claude Code marketplace plugin plus an explicit Codex plugin whitelist — командные стандарты, рабочие процессы, скиллы, хуки, and the first Codex-supported context-menu surface.
Что это даёт:
v2.0 BREAKING: npm install path удалён. Distribution через canonical Anthropic plugin marketplace (
/plugin marketplace add+/plugin install). Cursor support удалён полностью. Existing v1 users — см. Migration v1 → v2 ниже.
В Claude Code CLI или Desktop session:
/plugin marketplace add stgmt/dev-pomogator
/plugin install dev-pomogator@stgmt
/reload-plugins
В Claude Desktop: после /plugin install restart Desktop application для подхвата нового plugin (или используй UI «+» button → «Plugins» browser для install с visual feedback).
/plugin install dev-pomogator@stgmt --scope user # default — across all your projects
/plugin install dev-pomogator@stgmt --scope project # committed to <cwd>/.claude/settings.json (team-shared)
/plugin install dev-pomogator@stgmt --scope local # personal per-repo (gitignored settings.local.json)
Codex uses its own plugin marketplace command tree; there is no npm --codex install flag. From this checkout:
cd E:\repos\dev-pomogator
.\scripts\install-codex-context-menu.ps1
The script registers this checkout as a local Codex marketplace, installs context-menu@dev-pomogator-codex, and applies the Windows right-click menu with --codex-only.
This Codex path installs only Codex.nss, imports/Codex.nss, codex-icon.ico, and launch-Codex-tui.ps1; it does not create or overwrite Claude context-menu artifacts. The icon is extracted from the installed OpenAI Codex app when available; generated fallback is used only when the local app icon cannot be extracted. The current Codex entry is Codex (YOLO) with -NoTui; Codex+TUI is intentionally deferred until the TUI launcher path is verified.
После canonical install Claude Code copies plugin к ~/.claude/plugins/cache/stgmt/dev-pomogator/<version>/:
/dev-pomogator:create-spec, /dev-pomogator:run-tests, /dev-pomogator:plan-pomogator, /dev-pomogator:pomogator-doctor, /dev-pomogator:report-issue, /dev-pomogator:research-workflow и др. — invokable через Skill tool или slash command/reflect, /simplify, /pomogator-doctor, /report-issue — slash commands из plugin.claude-plugin/hooks.json plugin manifest.mcp.json.claude/rules/ content для context loadingenabledPlugins declaration автоматически добавляется в ~/.claude/settings.json соответствующего scope.
📍 Карта всех встроенных инструментов и связей между ними: docs/COMPONENTS.md — dev-pomogator это одна система (один плагин), внутри которой инструменты связаны контурами: спеки → валидация → согласованность, тесты → мониторинг → качество, statusline ↔ doctor и т.д.
Use /report-issue <description> (or /dev-pomogator:report-issue) to prepare a sanitized GitHub issue draft. The workflow shows the repository, title, body, and exact approval digest before it creates anything. On approval it uses GitHub CLI and opens the result URL. If GitHub CLI is missing, unauthenticated, or errors, it saves a Markdown draft and provides a filled GitHub new-issue URL instead. Authenticate with gh auth login to submit through GitHub CLI.
Часть плагина (не отдельный plugin): SessionStart-хук tools/native-statusline/ при первом
запуске сессии подключает ccstatusline как
основной statusline Claude Code и, если конфига виджетов ещё нет, создаёт
~/.config/ccstatusline/settings.json с 3-строчным столбиком (одна строка обрезается по
ширине терминала и съедает хвост):
Model: Opus 4.8 | Ctx: 232.0k
dev-pomogator | cwd: ~\dev-pomogator
⎇ feat/my-branch | (+555,-59)
Правила безопасности:
npx claudepluginhub stgmt/dev-pomogator --plugin dev-pomogatorHarness-native ECC operator layer - 67 agents, 278 skills, 94 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Lazy senior dev mode. Forces the simplest, shortest solution that actually works: YAGNI, stdlib first, no unrequested abstractions.
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Matt Pocock's agent skills for real engineering — grilling, spec/ticket flows, TDD, code review, domain modelling and more. Plug-and-play, not vibe coding.