dewil/ai-toolkit
AI Toolkit - агентская ОС для работы над проектами. Общая память, задачи и правила для AI-агентов разных вендоров.
stack: Python
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
Copy this rig
# review before running: this installs third-party code
$ npx degit dewil/ai-toolkit/.claude ./rig-ai-toolkit # inspect, then merge into .claude/ MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ npx degit dewil/ai-toolkit/skills/agent-browser .claude/skills/agent-browser $ npx degit dewil/ai-toolkit/skills/android-apk-release .claude/skills/android-apk-release $ npx degit dewil/ai-toolkit/skills/asana-blockers .claude/skills/asana-blockers $ npx degit dewil/ai-toolkit/skills/asana-comments .claude/skills/asana-comments $ npx degit dewil/ai-toolkit/skills/asana-project .claude/skills/asana-project $ npx degit dewil/ai-toolkit/skills/backlog-task .claude/skills/backlog-task $ npx degit dewil/ai-toolkit/skills/canon-sync .claude/skills/canon-sync $ npx degit dewil/ai-toolkit/skills/csv-xlsx .claude/skills/csv-xlsx $ npx degit dewil/ai-toolkit/skills/estimate-calibration .claude/skills/estimate-calibration $ npx degit dewil/ai-toolkit/skills/gdocs .claude/skills/gdocs $ npx degit dewil/ai-toolkit/skills/harvest-canon .claude/skills/harvest-canon $ npx degit dewil/ai-toolkit/skills/implement .claude/skills/implement $ npx degit dewil/ai-toolkit/skills/jev-decide .claude/skills/jev-decide $ npx degit dewil/ai-toolkit/skills/llm-ru .claude/skills/llm-ru $ npx degit dewil/ai-toolkit/skills/local-transcription .claude/skills/local-transcription $ npx degit dewil/ai-toolkit/skills/md-docx .claude/skills/md-docx $ npx degit dewil/ai-toolkit/skills/md-pdf .claude/skills/md-pdf $ npx degit dewil/ai-toolkit/skills/md-pptx .claude/skills/md-pptx $ npx degit dewil/ai-toolkit/skills/memory-compact .claude/skills/memory-compact $ npx degit dewil/ai-toolkit/skills/memory-search .claude/skills/memory-search $ npx degit dewil/ai-toolkit/skills/migrate .claude/skills/migrate $ npx degit dewil/ai-toolkit/skills/minimal-code .claude/skills/minimal-code $ npx degit dewil/ai-toolkit/skills/mymeet-snapshot .claude/skills/mymeet-snapshot $ npx degit dewil/ai-toolkit/skills/office-md .claude/skills/office-md $ npx degit dewil/ai-toolkit/skills/pii-mask .claude/skills/pii-mask $ npx degit dewil/ai-toolkit/skills/proofread .claude/skills/proofread $ npx degit dewil/ai-toolkit/skills/qa-browser .claude/skills/qa-browser $ npx degit dewil/ai-toolkit/skills/redmine-snapshot .claude/skills/redmine-snapshot $ npx degit dewil/ai-toolkit/skills/repo-publication .claude/skills/repo-publication $ npx degit dewil/ai-toolkit/skills/requirements .claude/skills/requirements $ npx degit dewil/ai-toolkit/skills/review .claude/skills/review $ npx degit dewil/ai-toolkit/skills/sdd .claude/skills/sdd $ npx degit dewil/ai-toolkit/skills/session-cost .claude/skills/session-cost $ npx degit dewil/ai-toolkit/skills/telegram-send .claude/skills/telegram-send $ npx degit dewil/ai-toolkit/skills/telegram-snapshot .claude/skills/telegram-snapshot $ npx degit dewil/ai-toolkit/skills/tutor .claude/skills/tutor $ npx degit dewil/ai-toolkit/skills/web-research .claude/skills/web-research
$ curl -fsSL --create-dirs -o .claude/agents/prompt-reviewer.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/.claude/agents/prompt-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/architect.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/architect.md $ curl -fsSL --create-dirs -o .claude/agents/code-reviewer.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/code-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/copy-editor.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/copy-editor.md $ curl -fsSL --create-dirs -o .claude/agents/debugger.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/debugger.md $ curl -fsSL --create-dirs -o .claude/agents/explorer.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/explorer.md $ curl -fsSL --create-dirs -o .claude/agents/implementer.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/implementer.md $ curl -fsSL --create-dirs -o .claude/agents/librarian.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/librarian.md $ curl -fsSL --create-dirs -o .claude/agents/note-taker.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/note-taker.md $ curl -fsSL --create-dirs -o .claude/agents/note-writer.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/note-writer.md $ curl -fsSL --create-dirs -o .claude/agents/planner.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/planner.md $ curl -fsSL --create-dirs -o .claude/agents/searcher.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/searcher.md $ curl -fsSL --create-dirs -o .claude/agents/summarizer.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/summarizer.md $ curl -fsSL --create-dirs -o .claude/agents/test-writer.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/test-writer.md $ curl -fsSL --create-dirs -o .claude/agents/tracker.md https://raw.githubusercontent.com/dewil/ai-toolkit/main/agents/tracker.md
This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 16.2k rigs:
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(./.env.*)",
"Read(**/.env)",
"Read(~/.ssh/**)",
"Bash(git push --force:*)",
"Read(**/*.pem)",
"Read(.env)",
"Read(~/.aws/**)",
"Read(**/*.key)",
"Read(**/.env.*)",
"Bash(git push --force*)",
"Bash(rm -rf:*)",
"Bash(rm -rf *)",
"Bash(sudo *)",
"Bash(sudo:*)",
"Bash(rm -rf /)",
"Bash(git push -f:*)",
"Read(.env.*)",
"Bash(git reset --hard*)",
"Bash(rm -rf /*)"
],
"ask": [
"Bash(git push:*)",
"Bash(git push *)",
"Bash(gh pr merge *)",
"Bash(git commit:*)",
"Bash(rm *)",
"Bash(gh pr merge:*)",
"Bash(gh pr create *)",
"Bash(git reset --hard:*)",
"Bash(npm publish:*)",
"Bash(git rebase *)"
]
}
} Skills (37)
agent-browserandroid-apk-releaseasana-blockersasana-commentsasana-projectbacklog-taskcanon-synccsv-xlsxestimate-calibrationgdocsharvest-canonimplementjev-decidellm-rulocal-transcriptionmd-docxmd-pdfmd-pptxmemory-compactmemory-searchmigrateminimal-codemymeet-snapshotoffice-mdpii-maskproofreadqa-browserredmine-snapshotrepo-publicationrequirementsreviewsddsession-costtelegram-sendtelegram-snapshottutorweb-research
Subagents (15)
| prompt-reviewer model: sonnet | Reviews prompts, agent definitions, and rule files for clarity, safety, and idempotency. Use proactively immediately after any prompt or canon file is written or modified. |
| architect model: opus | Designs system architecture, evaluates trade-offs, and plans complex refactors. Use when the task requires deep reasoning about structure or design decisions. |
| code-reviewer model: sonnet | Reviews code for quality, security, and best practices. Use proactively immediately after any code is written or modified. |
| copy-editor model: sonnet | Reviews documentation and prose for readability, consistency, and clarity. Use proactively immediately after any documentation is written or modified. |
| debugger model: sonnet | Analyzes errors, test failures, and unexpected behavior. Use proactively when any error, exception, or failing test is encountered. |
| explorer model: haiku | Explores project structure, reads documentation, and maps architecture. Use proactively when needing to understand how modules are connected. |
| implementer model: sonnet | Реализация по готовой спеке и независимым тестам. Применяет изменения в выделенной области, возвращает проверки и отклонения. |
| librarian model: haiku | Ревизует Obsidian-vault на гигиену - битые [[ссылки]], дубли заметок по близким названиям, сироты без входящих ссылок, заметки без тегов. Read-only, отчетом. Использовать проактивно после серии правок |
| note-taker model: sonnet | Converts lecture transcripts into structured study notes with key concepts, definitions, examples, and open questions. Use when given a transcript or recording text from a lecture. |
| note-writer model: sonnet | Создает и правит заметки в Obsidian-vault'е по конвенциям проекта (wiki-notes-style, wiki-linking-obsidian, wiki-conventions). Перед созданием новой заметки ищет существующую близкую по теме и предлаг |
| planner model: opus | Decomposes goals into a minimal sequence of verifiable steps with owners, dependencies, and explicit risks. Use when the user describes an outcome (a feature, a refactor, a research goal, a process ch |
| searcher model: haiku | Searches files, patterns, symbols, and dependencies in the codebase. Use proactively for any grep, glob, or file discovery task before making changes. |
| summarizer model: sonnet | Compresses long input (transcripts, threads, documents, lectures, meeting notes) into a structured summary with key points, decisions or takeaways, and open questions. Use when given a long source to |
| test-writer model: sonnet | Независимые тесты по спеке и публичному контракту без чтения реализации. Для SDD получает чистый контекст и возвращает RED, критерии и вопросы. |
| tracker model: sonnet | Tracks status of tasks, owners, deadlines, and blockers across plans and meeting notes. Use proactively to surface what's overdue, who's blocked, and what hasn't been touched recently. |
Slash commands (3)
/check-inbox/canon/feedback-inbox
Permissions
deny (0)
—
ask (0)
—
allow (4)
Skill(update-config)
Skill(update-config:*)
Bash(git push origin main)
Bash(git pull *)
Similar rigs
Yassinekrn/claude-setup
Personal Claude Code setup — orchestrator/subagent delegation policy, custom agents, and a colored PowerShell statusline (git branch, model+effort, tokens, cost, rate limits).
Orchestrator 1.2k tok · · Q15
kapadias/nonna
Your AI agent says done. Nonna makes it prove it: she runs your test suite before a coding agent can stop, and blocks pushes to main and secrets in files. Claude Code plugin, plus git hooks for Codex, Cursor, Copilot, Gemini and more
Fort Knox 3.9k tok · · Q78
yassine-krichen/claude-code-agentic-setup
Personal Claude Code setup — orchestrator/subagent delegation policy, custom agents, and a colored PowerShell statusline (git branch, model+effort, tokens, cost, rate limits).
Orchestrator 2.3k tok · · Q23
theishandubey/claude-config
A complete Claude Code setup. An agent roster, an orchestration playbook, safe defaults and curated skills, installed with one command.
Orchestrator 5.0k tok · · Q5