RicardoSantos0/claudius
A governed multi-agent delivery framework for Claude Code: reusable agents, MAS workflow skills, policies, templates, and a Python CLI for project lifecycle, handoffs, shared state, and evaluation.
ARCHETYPE
YOLO Cowboy
Permissions? Never heard of 'em. Ships at the speed of `--dangerously-skip-permissions`.
Copy this rig
# review before running: this installs third-party code $ claude mcp add mas -e UV_LINK_MODE=YOUR_UV_LINK_MODE -- uv run --extra server mas-server $ npx degit RicardoSantos0/claudius/.claude ./rig-claudius # 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.
$ claude mcp add mas -e UV_LINK_MODE=YOUR_UV_LINK_MODE -- uv run --extra server mas-server $ npx degit RicardoSantos0/claudius/skills/frontend-design .claude/skills/frontend-design $ npx degit RicardoSantos0/claudius/skills/graphify .claude/skills/graphify $ npx degit RicardoSantos0/claudius/skills/mas-clarify .claude/skills/mas-clarify $ npx degit RicardoSantos0/claudius/skills/mas-document .claude/skills/mas-document $ npx degit RicardoSantos0/claudius/skills/mas-examine .claude/skills/mas-examine $ npx degit RicardoSantos0/claudius/skills/mas-handoff .claude/skills/mas-handoff $ npx degit RicardoSantos0/claudius/skills/mas-logwork .claude/skills/mas-logwork $ npx degit RicardoSantos0/claudius/skills/mas-plan .claude/skills/mas-plan $ npx degit RicardoSantos0/claudius/skills/mas-postmortem .claude/skills/mas-postmortem $ npx degit RicardoSantos0/claudius/skills/mas-review .claude/skills/mas-review $ npx degit RicardoSantos0/claudius/skills/skill-builder .claude/skills/skill-builder
$ curl -fsSL --create-dirs -o .claude/agents/_utilities.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/_utilities.md $ curl -fsSL --create-dirs -o .claude/agents/devils_advocate.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/devils_advocate.md $ curl -fsSL --create-dirs -o .claude/agents/domain_expert.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/domain_expert.md $ curl -fsSL --create-dirs -o .claude/agents/efficiency_advisor.md https://raw.githubusercontent.<redacted>_advisor.md $ curl -fsSL --create-dirs -o .claude/agents/evaluator_agent.md https://raw.githubusercontent.<redacted>_agent.md $ curl -fsSL --create-dirs -o .claude/agents/hr_agent.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/hr_agent.md $ curl -fsSL --create-dirs -o .claude/agents/inquirer_agent.md https://raw.githubusercontent.<redacted>_agent.md $ curl -fsSL --create-dirs -o .claude/agents/master_orchestrator.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/master_orchestrator.md $ curl -fsSL --create-dirs -o .claude/agents/product_manager_agent.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/product_manager_agent.md $ curl -fsSL --create-dirs -o .claude/agents/project_manager_agent.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/project_manager_agent.md $ curl -fsSL --create-dirs -o .claude/agents/quality_advisor.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/quality_advisor.md $ curl -fsSL --create-dirs -o .claude/agents/reliability_engineer.md https://raw.githubusercontent.<redacted>_engineer.md $ curl -fsSL --create-dirs -o .claude/agents/risk_advisor.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/risk_advisor.md $ curl -fsSL --create-dirs -o .claude/agents/scribe_agent.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/scribe_agent.md $ curl -fsSL --create-dirs -o .claude/agents/session_scheduler.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/session_scheduler.md $ curl -fsSL --create-dirs -o .claude/agents/spawner_agent.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/spawner_agent.md $ curl -fsSL --create-dirs -o .claude/agents/trainer_agent.md https://raw.githubusercontent.com/RicardoSantos0/claudius/main/agents/trainer_agent.md
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(./.env.*)",
"Read(./secrets/**)",
"Read(./skills/notebooklm/data/browser_state/**)",
"Read(./skills/notebooklm/data/auth_info.json)",
"Read(./mas/data/**)",
"Read(./mas/projects/**)",
"Read(./.git/**)",
"Bash(curl *)",
"Bash(wget *)",
"Bash(rm -rf *)"
]
}
} MCP servers (1)
| server | source | est. tokens |
|---|---|---|
| server · "mas" env: UV_LINK_MODE | pypi | 2.5k |
Skills (11)
frontend-designgraphifymas-clarifymas-documentmas-examinemas-handoffmas-logworkmas-planmas-postmortemmas-reviewskill-builder
Subagents (17)
| _utilities | — |
| devils-advocate model: inherit | Devil's Advocate on the Master's Consultant Panel. Constructively challenges assumptions, proposes alternative perspectives, identifies hidden incentive problems, and questions consensus that forms to |
| domain-expert model: inherit | Domain Expert on the Master's Consultant Panel. Applies deep domain knowledge to every question — best practices, prior art, domain-specific constraints and risks. Prompt is dynamically enriched with |
| efficiency-advisor model: inherit | Efficiency Advisor on the Master's Consultant Panel. Views every decision through the lens of simplicity, resource efficiency, and overhead minimization. Flags overengineering, estimates costs, and id |
| evaluator-agent model: inherit | Performance Evaluation Agent of the Governed Multi-Agent Delivery System. Invoked automatically after project completion (or manually by the Master Orchestrator). Collects project data, scores metrics |
| hr-agent model: inherit | HR Agent of the Governed Multi-Agent Delivery System. Invoked by the Master Orchestrator to discover existing capabilities, evaluate matches, produce Deployment Recommendations for each capability nee |
| inquirer-agent model: inherit | Inquirer Agent of the Governed Multi-Agent Delivery System. Invoked by the Master Orchestrator to conduct structured intake of raw project briefs: analyze completeness, ask targeted clarification ques |
| master-orchestrator model: inherit | Master Orchestrator of the Governed Multi-Agent Delivery System. Invoke when coordinating a full project lifecycle: intake, planning, capability discovery, execution, evaluation, and improvement. Owns |
| product-manager-agent model: inherit | Product Manager Agent of the Governed Multi-Agent Delivery System. Invoked by the Master Orchestrator after a clarified specification is ready. Produces a structured product plan: goals, requirements |
| project-manager-agent model: inherit | Project Manager Agent of the Governed Multi-Agent Delivery System. Invoked by the Master Orchestrator after a product plan is approved. Decomposes scope into milestones and tasks, maps dependencies, r |
| quality-advisor model: inherit | Quality Advisor on the Master's Consultant Panel. Invoked by the Master Orchestrator to review decisions for completeness, measurability, testability, and quality standards. Flags vague criteria, miss |
| reliability-engineer model: inherit | Python delivery agent for quality gates and test infrastructure. Owns test suite (>=80% coverage enforced), golden fixtures, CI lint guards, write-path interception tests at transport layer, packaging |
| risk-advisor model: inherit | Risk Advisor on the Master's Consultant Panel. Invoked by the Master Orchestrator to analyze risk for significant decisions. Views every question through failure modes, blast radius, safeguards, and r |
| scribe-agent model: inherit | Scribe Agent of the Governed Multi-Agent Delivery System. Invoked by the Master Orchestrator to create and maintain durable project memory: project folders, decision logs, handoff records, artifact re |
| session-scheduler model: inherit | Scheduled session-resume agent. Checks for active MAS projects with incomplete work, acquires a per-project lock to prevent duplicate runs, then invokes /resume-mas to continue the project from its la |
| spawner-agent model: inherit | Agent Designer of the Governed Multi-Agent Delivery System. Invoked only after Master Orchestrator approves a gap certificate. Validates all spawn policy rules, then produces a draft agent package (de |
| trainer-agent model: inherit | Improvement Proposal Agent of the Governed Multi-Agent Delivery System. Invoked after every evaluation cycle. Reads evaluation reports, identifies improvement patterns, and produces advisory proposals |
Slash commands (2)
/markitdown/resume-mas
Permissions
deny (11)
Read(./.env)
Read(./.env.*)
Read(./secrets/**)
Read(./skills/notebooklm/data/browser_state/**)
Read(./skills/notebooklm/data/auth_info.json)
Read(./mas/data/**)
Read(./mas/projects/**)
Read(./.git/**)
Bash(curl *)
Bash(wget *)
Bash(rm -rf *)
ask (0)
—
allow (6)
Bash(*)
Read(*)
Write(*)
Edit(*)
Glob(*)
Grep(*)
Similar rigs
quboqin/template-agenticide-vibecoding
🤖 Claude Code定制化研发流程系统 - 基于Spec-Driven和Test-Driven的AI智能体团队协作框架。构建AI优先的软件开发新范式，让AI智能体团队替代传统人类开发团队。
Orchestrator 1.6k tok ·
xahdiyar37-hub/template-agenticide-vibecoding
Claude Code agentic AI development workflow template - 7 AI Agents, spec-driven + test-driven development
Orchestrator 1.6k tok ·
yusufkaraaslan/Skill_Seekers
Convert documentation websites, GitHub repositories, and PDFs into Claude AI skills with automatic conflict detection
Pragmatist 8.1k tok ·
mjacobus/dotfiles
My .dotfiles
Minimalist 142 tok ·