radif-ru/ai-multi-agent-system
Локальная мульти-агентная система на self-hosted LLM (Ollama), 0 обращений к облачным LLM-API. Роли Planner/Executor/Critic, цикл thought → action → observation. 3 канала: Telegram, MAX, консоль. 19 инструментов: почта, Яндекс.Диск, веб, OC
ARCHETYPE
Skill Collector
Ten-plus skills loaded on demand. A procedural-knowledge library.
Copy this rig
$ git clone --depth 1 https://github.com/radif-ru/ai-multi-agent-system 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 radif-ru/ai-multi-agent-system/.agents/skills/architecture-discipline .claude/skills/architecture-discipline $ npx degit radif-ru/ai-multi-agent-system/.agents/skills/async-discipline .claude/skills/async-discipline $ npx degit radif-ru/ai-multi-agent-system/.agents/skills/automation-discipline .claude/skills/automation-discipline $ npx degit radif-ru/ai-multi-agent-system/.agents/skills/debugging-discipline .claude/skills/debugging-discipline $ npx degit radif-ru/ai-multi-agent-system/.agents/skills/documentation-discipline .claude/skills/documentation-discipline $ npx degit radif-ru/ai-multi-agent-system/.agents/skills/error-handling-discipline .claude/skills/error-handling-discipline $ npx degit radif-ru/ai-multi-agent-system/.agents/skills/git-discipline .claude/skills/git-discipline $ npx degit radif-ru/ai-multi-agent-system/.agents/skills/pr-review-discipline .claude/skills/pr-review-discipline $ npx degit radif-ru/ai-multi-agent-system/.agents/skills/prompt-injection-defense .claude/skills/prompt-injection-defense $ npx degit radif-ru/ai-multi-agent-system/.agents/skills/pull-request-discipline .claude/skills/pull-request-discipline $ npx degit radif-ru/ai-multi-agent-system/.agents/skills/skill-authoring .claude/skills/skill-authoring $ npx degit radif-ru/ai-multi-agent-system/.agents/skills/testing-discipline .claude/skills/testing-discipline $ npx degit radif-ru/ai-multi-agent-system/app/skills/about-project .claude/skills/about-project $ npx degit radif-ru/ai-multi-agent-system/app/skills/bash-linux .claude/skills/bash-linux $ npx degit radif-ru/ai-multi-agent-system/app/skills/bash-pro .claude/skills/bash-pro $ npx degit radif-ru/ai-multi-agent-system/app/skills/code-review .claude/skills/code-review $ npx degit radif-ru/ai-multi-agent-system/app/skills/email-assistant .claude/skills/email-assistant $ npx degit radif-ru/ai-multi-agent-system/app/skills/email_draft .claude/skills/email_draft $ npx degit radif-ru/ai-multi-agent-system/app/skills/example-summary .claude/skills/example-summary $ npx degit radif-ru/ai-multi-agent-system/app/skills/humanizer .claude/skills/humanizer $ npx degit radif-ru/ai-multi-agent-system/app/skills/mail-digest .claude/skills/mail-digest $ npx degit radif-ru/ai-multi-agent-system/app/skills/ocr .claude/skills/ocr $ npx degit radif-ru/ai-multi-agent-system/app/skills/scheduler .claude/skills/scheduler $ npx degit radif-ru/ai-multi-agent-system/app/skills/weather .claude/skills/weather
This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 7,204 rigs:
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(**/.env)",
"Read(~/.ssh/**)",
"Bash(rm -rf *)",
"Read(**/*.pem)",
"Bash(rm -rf /)",
"Bash(git push --force:*)",
"Bash(sudo *)",
"Read(.env)",
"Bash(rm -rf /*)",
"Read(./.env.*)",
"Read(~/.aws/**)",
"Bash(git push --force*)",
"Bash(rm -rf:*)",
"Read(**/*.key)",
"Read(**/.env.*)",
"Bash(sudo:*)",
"Bash(git reset --hard*)",
"Bash(git reset --hard:*)",
"Read(.env.*)"
],
"ask": [
"Bash(git push:*)",
"Bash(git push *)",
"Bash(git commit:*)",
"Bash(rm *)",
"Bash(rm:*)",
"Bash(git rebase *)",
"Bash(wget *)",
"Bash(npm publish:*)",
"Bash(git commit *)",
"Bash(gh pr merge *)"
]
}
} Skills (24)
architecture-disciplineasync-disciplineautomation-disciplinedebugging-disciplinedocumentation-disciplineerror-handling-disciplinegit-disciplinepr-review-disciplineprompt-injection-defensepull-request-disciplineskill-authoringtesting-disciplineabout-projectbash-linuxbash-procode-reviewemail-assistantemail_draftexample-summaryhumanizermail-digestocrschedulerweather
Cursor rules (1)
agents · always
Similar rigs
bzoran/ai-agent-patterns-and-workflow
AI agent patterns and workflow templates for Claude Code - skills, hooks, agents, and ticket-driven development
Pragmatist 803 tok ·
kolega-ai/kolega-code
Agentic coding in the terminal: the model writes its own multi-agent workflows (Gigacode). Provider-agnostic, local-first, 15+ model providers, MCP support.
Pragmatist 601 tok ·
timwuhaotian/the-pair
Open-source AI pair programming for desktop: a Mentor + Executor agent cross-check each other's code to catch AI hallucinations. Works with Claude Code, Codex, Gemini & opencode. macOS / Windows / Linux.
Pragmatist 5.5k tok ·
Aboudjem/humanizer-skill
Free, open-source AI writing humanizer and detector. 55 patterns, 5 voices, a 0-100 AI-tell score, and nothing leaves your machine.
Minimalist 1.7k tok ·