Mega-Mind Skills System v1.1.0

A unified Gold Standard skill system for any AI coding assistant — covering the full software development lifecycle from planning through production operations and compliance.

53 Gold Standard Skills — Core workflow, domain expert, meta/learning, and token optimization — every skill meets the Gold Standard SKILL.md v2.0 specification with 12 required sections including Blocking Violations, Verification with quality gates, Performance & Cost models, and Anti-Patterns tables. Compliance is machine-verified in CI — the routing matrix, skills listing, and every catalog table are generated from skills-manifest.json + routing.json and can never drift from the skills themselves.
11 Agent Personas — Deep-dive personas for architecture, QA, security, accessibility, adversarial testing, privacy, incidents, and releases.
9 Workflow Chains — Executable sequences covering the full lifecycle: brainstorm → plan → build → verify → ship → release → operate → learn.
Multi-Platform — Works with Antigravity, Claude Code, GitHub Copilot, Cursor, OpenCode, Codex, and pi — any tool supporting the Agent Skills open standard.

New in v1.1.0 (2026-11-11): The routing layer is now executable and tested. A deterministic router (mega_mind/router.py) backs a new mmo route "<request>" command, applying the precedence the orchestrator only documented before — no model call. tests/test_router.py proves 78/78 curated phrases, 53/53 skills, 11/11 agents and the skill's own examples route to their documented target and layer. Writing the tests exposed four real defects (whole-chain requests unrouted, published examples matching nothing, a CLI crash on non-UTF-8 consoles, routes not kind-checked). See the changelog for details.

New in v1.0.5 (2026-10-14): All 11 agent personas were prompt-tested against their own documented rules and hardened where the audit found gaps (blast-radius approval, ADR cost and override rules, a go/no-go release gate, DSAR and lawful basis, incident comms cadence, delegation and estimate discipline). Routing is consistent across both layers — 57 routes, every matrix phrase also a frontmatter trigger — plus a new agent_refs validator check for stale persona links. See the changelog for details.

New in v1.0 (2026-07-09): All 53 skills upgraded to Gold Standard. 19,531 total lines, 12/12 sections compliance. See the changelog for details.

Quick Start

# Install the CLI
pip install mmo          # or
pipx install mmo         # (recommended) or
uv tool install mmo      # or
uvx mmo                  # run without installing

# Initialize in your project
cd /path/to/your/project
uvx mmo init             # creates .agent/ with all 53 skills

Then in your AI assistant:

/mega-mind I need to add user authentication
/mega-mind route "fix this production bug"
/brainstorm How should I design the payment flow?
/verify

Read the full installation guide →

What's Inside

Orchestrator

The /mega-mind command is your primary entry point. It analyzes requests, routes to the appropriate skill or agent, chains multi-step workflows, and tracks progress via task.md.

Category	Count	Purpose
Core Workflow	12	Disciplined development practices: planning, execution, review, verification
Domain Expert	29	Specialized technical expertise across architecture, DevOps, data, security
Meta & Learning	8	Efficiency and continuous improvement loops
Token Optimization	4	Token reduction and context management
Total	53

Agent Personas

Category	Personas
Development	`tech-lead`, `planner`, `architect`
Quality & Testing	`code-reviewer`, `qa-engineer`, `accessibility-auditor`, `adversarial-tester`
Security & Compliance	`security-reviewer`, `data-privacy-officer`
Operations & Releases	`incident-commander`, `release-manager`

Workflow Chains

Workflow	Purpose
brainstorm	Structured exploration
write-plan	Implementation planning
execute-plan	Disciplined step-by-step execution
high-complexity-dev	Multi-agent orchestration
review	Structured code review
debug	Root cause analysis
ship	Merge, deploy, branch cleanup
incident-response	Production incident lifecycle
release	Versioning, rollout, monitoring

Full reference →

Compatibility

Any AI coding agent (Antigravity, Claude Code, Copilot, Cursor, OpenCode, Codex, pi, and all tools supporting the Agent Skills open standard).