Leo — AI skill for Claude Code
A deterministic context-driven agentic SDLC framework for Cursor and Claude.
How to install Leo
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open alex-zaporozhan/leo and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Leo does
A deterministic context-driven agentic SDLC framework for Cursor and Claude. 41 codified laws and 22 specialist roles to eliminate LLM hallucinations and context drift.
Alternatives in AI
- The Pair — Open-source AI pair programming for desktop: a Mentor + Executor agent cross-check each other's code to catch 358 ★
- SoloDawn — Catches and repairs AI hallucinations before delivery: three quality gates, 31 built-in rules, a self-healing 312 ★
- Iran MCP — 39 MCP servers (488 tools) that give AI agents the Iranian web: Digikala, Torob, Divar, Snappfood, Jabama, MrB 151 ★
README
LEO
Lead Engineering Orchestrator
**A written engineering process for coding agents.** Markdown you load instead of a one-line personality. Nothing installs.
[](./LICENSE) [](./CASE_STUDIES.md) [](#compatibility)
[In brief](#in-brief) · [Sixty-second test](#sixty-second-test-before-you-read-any-further) · [What it looks like](#what-it-actually-looks-like) · [How it works](#how-it-works) · [Proof at scale](#proof-at-scale) · [Get started](#get-started) · [Business case](./BUSINESS_CASE.md) · [Steering the frontend](./GUIDE_FRONTEND_CONTROL.md) · [Manifesto](./MANIFESTO.md) · [License](#license)
In brief
Coding agents fail in four repeatable ways: they forget a decision made forty messages ago and re-make it differently; they skip the expensive 20% — the empty state, the error contract, the timeout on the outgoing call — because nothing made skipping it cost anything; they report "should work now" in the same tone as a verified fact; and nothing in a single-agent chat ever tells them no.
LEO is that missing process, written into the repository as rules the agent is loaded under:
- A router. Every task resolves to one of 22 classes before anything is read; the class names the ≤6 files to open, in order. The cap does not grow with the library, so a 133-file library costs a landing page nothing it does not use.
- Numbered laws, a precedence ladder, a conflict registry. 44 rules, most with an incident behind them. A reviewing pass can cite Law 11; it cannot cite "be careful with async." Where two laws collide, the collision is decided once and written down.
- **Reflexes, not self-rev
Related Skills
Agent Context Framework
Production-ready Markdown (.md) context templates for AI coding agents (Cursor, Claude Code, Aider, Windsurf).
Agent Guides
Install a full engineering team into your AI agent. 7 SDLC roles · 105+ skills · 18 areas — Claude Code, openc
Tribunal Kit
A universal, native-Rust governance layer for AI coding agents. Enforces Subagent-Driven Development (SDD) to
Developing Llamaindex Systems
Production-grade LlamaIndex agentic system skill for Claude Code. Covers semantic ingestion, property graphs,
AI Engineering Claude C1 Harness Engineering Project
Four Claude Code systems from the Harness Engineering capstone: a stop_reason-driven agentic loop with tool de
DCGP
Context governance kernel for LLM agents. Predicts entropy, blocks hallucinations. Pairs with opencode-dcp. ll
Related Agents
Agentic Auditor
PYA specialist spawned by the /pya orchestrator. Audits AI agent and LLM surfaces in an untrusted repository:
RAG Coder
Context-aware coder that grounds all code generation in actual codebase using RAG to eliminate hallucinations
Spec Agentic AI Solutions Engineer
AI solutions engineer who builds AI agents and AI-driven applications. Use this agent proactively when designi