llms.txt · git:20261020.d5cf377 · 2026-10-20 · sha256 57d6392d441ecfc8
llms.txt git:20261020.d5cf377A
Immutable. This exact content never changes and is served at /api/v1/blob/57d6392d441ecfc8.
# Agentic OS > Agentic OS is an open-source governance framework for AI coding agents. It gives agents — Claude Code, OpenAI Codex, Cursor, GitHub Copilot, Google Antigravity, and any Markdown-reading LLM agent — a repeatable plan → build → review → test → ship workflow with quality gates: a credential scan and a validator that fails when a task's work trail is missing a phase or its evidence. Agentic OS installs into any project as a set of Markdown governance files plus optional Python validation tooling. Every task is classified (tiny-fix, quick-win, feature, hotfix, architecture-change) and flows through mandatory phases; each phase is expected to leave a record and its evidence in the task's work trail, which a local validator checks; and a single source of truth (SSoT) persists decisions across agent sessions so context survives between conversations. ## Documentation - [README](https://raw.githubusercontent.com/KbWen/agentic-os/main/README.md): Overview, install, features, and FAQ - [Quick Start](https://github.com/KbWen/agentic-os/blob/main/README.md#quick-start): Install commands and first steps - [Commands](https://github.com/KbWen/agentic-os/blob/main/docs/reference.md#commands): The workflow command reference - [Agent Philosophy](https://raw.githubusercontent.com/KbWen/agentic-os/main/.agentcortex/docs/AGENT_PHILOSOPHY.md): The 10 core principles - [Model Selection Guide](https://raw.githubusercontent.com/KbWen/agentic-os/main/docs/AGENT_MODEL_GUIDE.md): Choosing an AI model - [Lifecycle Benchmark](https://raw.githubusercontent.com/KbWen/agentic-os/main/docs/LIFECYCLE_BENCHMARK.md): Real token-cost data across 6 scenarios - [Testing Protocol](https://raw.githubusercontent.com/KbWen/agentic-os/main/.agentcortex/docs/TESTING_PROTOCOL.md): Evidence and verification standards ## Platforms - Claude Code — native, `CLAUDE.md` entry point - OpenAI Codex — native, `AGENTS.md` entry point - Google Antigravity — native, `GEMINI.md` entry point + intent router - Cursor — compatible, `AGENTS.md` / project rules - GitHub Copilot — compatible, repository instructions - Any LLM agent — compatible, model-agnostic Markdown workflows ## License - MIT