etyb.ai  ·  v2.1.0  ·  Changelog  ·  Install Guide  ·  Architecture

etyb-skills

Install a virtual engineering company. Your AI agent gets a CTO, 20 specialists, and 9 always-on engineering disciplines — out of the box.

30 coordinated skills that turn any agentskills.io-compliant agent into a team that ships features methodically: brainstorming before architecture, tests before code, reviews before commits, evidence before "done."

Quick Start

# Claude Code — native plugin
/plugin marketplace add e-t-y-b/etyb-skills
/plugin install etyb-full@etyb-skills

# OpenAI Codex, Google Antigravity, or manual install
git clone https://github.com/e-t-y-b/etyb-skills.git
./etyb-skills/scripts/install.sh

# Codex projects — add runtime hooks + custom agents
./etyb-skills/scripts/install-codex-runtime.sh --target /path/to/your-project

Platform Support

Who This Is For

• Teams shipping real software — your AI agent works like a disciplined team member, not a solo cowboy
• Engineers tired of AI "yes-and" behavior — agents that skip tests, rubber-stamp reviews, or chase symptoms
• Regulated or high-stakes codebases (fintech, healthcare, e-commerce) — traceable decisions, gated releases, evidence-backed claims
• Solo developers — every specialist works standalone; the CTO is optional

What You Get

An AI coding agent that works like a 100-person engineering org:

• Refuses to ship untested code — TDD enforcement with deterministic hooks, not just instructions
• Stops you from building the wrong thing — structured brainstorming before architecture
• Pushes back on bad review feedback — evaluates findings on merit, no performative agreement
• Coordinates parallel work — subagent dispatch with two-stage review and worktree isolation
• Covers the full SDLC — from research through production operations
• Knows your domain — fintech ledgers, HIPAA compliance, e-commerce patterns, real-time systems

Architecture

USER REQUEST
     |
ETYB (CTO — routes, enforces gates, tracks plans)
     |
     |  ALWAYS-ON PROTOCOL LAYER
     |  |-- TDD — no code without failing test
     |  |-- Verification — evidence before claims
     |  |-- Review — no performative agreement
     |  |-- Plan execution — one task at a time
     |  |-- Brainstorm-first — explore before solving
     |  |-- Branch safety — never merge without green tests
     |  |-- Subagent coordination — parallel dispatch + review
     |  |-- Self-improvement — failing eval before skill changes
     |  +-- Debugging — root-cause-first after repeated failures
     |
DOMAIN EXPERTS (14 core teams + 6 vertical specialists)
     |
DEEP REFERENCES (100+ files, loaded on demand)

The 30 Skills

ETYB (1)

Your virtual CTO. Routes requests, enforces 5-phase gates (Design > Plan > Implement > Verify > Ship), mandates experts, tracks living plans. Works standalone, or as the conductor that pulls the other 29 skills into a coordinated team.

Core Team (14)

Domain Specialists (6)

Process Protocols (9)

Install

Claude Code (plugin)

# Everything
/plugin marketplace add e-t-y-b/etyb-skills
/plugin install etyb-full@etyb-skills

# Or install subsets
/plugin install etyb-process-protocols@etyb-skills   # 9 protocols + ETYB
/plugin install etyb-core-team@etyb-skills            # 14 core teams + ETYB
/plugin install etyb-verticals@etyb-skills            # 6 domain specialists

Codex, Antigravity, or Manual

git clone https://github.com/e-t-y-b/etyb-skills.git
cd etyb-skills
./scripts/install.sh               # detects platform, installs skills
./scripts/install.sh --dry-run     # preview without writing

Codex Runtime (hooks + agents)

./scripts/install-codex-runtime.sh --target /path/to/your-project

Installs .codex/config.toml, lifecycle hooks, and 4 custom agents (explorer, planner, reviewer, docs researcher). Backs up existing .codex/ on conflict. See docs/installation.md for the full guide.

Updating

./scripts/update.sh --check   # is there a newer version?
./scripts/update.sh           # interactive update (shows before/after)
./scripts/update.sh --force   # skip confirmation prompts

Preserves .etyb/plans/, .claude/plans/, and .claude/settings.local.json. Uses git merge --ff-only — no destructive operations. See CHANGELOG.md for version history.

How Skills Load (Token Efficiency)

Skills use progressive disclosure — a markdown-based RAG pattern:

Layer 0  Runtime guardrails (0 tokens — scripts outside the LLM)
Layer 1  ETYB always loaded (~3,500 tokens — the culture)
Layer 2  Relevant skill loads on demand (~2,500 tokens — the router)
Layer 3  Single reference loads on demand (~4,000 tokens — deep knowledge)

Per-activation: ~6,000-10,000 tokens (not the whole system)
Tier 0-1 requests: 0 extra tokens (ETYB handles directly)

Evidence: With vs Without Skills

Average token overhead: +53%. What you get: an engineer who says "no" when it matters.

Links

• Website: etyb.ai
• Repo: github.com/e-t-y-b/etyb-skills
• Latest release: v2.1.0
• Changelog: CHANGELOG.md
• Install guide: docs/installation.md
• Architecture: docs/architecture.md
• Issues & contributions: github.com/e-t-y-b/etyb-skills/issues

License

MIT — see LICENSE.