Cross-Platform Transfer

Stato stores expertise as Python files in .stato/. Bridge files are lightweight pointers (~500 tokens) that tell each platform’s agent where to find and how to use those modules.

Supported Platforms

Platform	Bridge File	Auto-Read	Command
Claude Code	CLAUDE.md	Yes	stato bridge --platform claude
Cursor	.cursorrules	Yes	stato bridge --platform cursor
Codex	AGENTS.md	Yes	stato bridge --platform codex
Generic	README.stato.md	No (manual)	stato bridge --platform generic

Generate Bridges

Generate a bridge for one platform:

stato bridge --platform claude

Generate all bridges at once:

stato bridge --platform all

Auto-detect (defaults to Claude Code):

stato bridge --platform auto

What a Bridge Contains

Each bridge file includes:

1. Header: identifies the project as a stato-managed project
2. Module index: table of all modules with their type, class, and path
3. On-demand loading instructions: tells the agent to read full skill files only when the skill is relevant to the current task
4. Working rules: guidelines for the agent (read before coding, update memory after milestones, use stato resume after context loss, etc.)

The bridge is a table of contents, not a copy of the modules. This keeps token usage low (~500 tokens regardless of how many modules you have) while giving the agent everything it needs to locate and load expertise on demand.

Same Expertise, Every Platform

The key insight is that .stato/ is platform-independent. The Python module files are the same whether you use Claude Code, Cursor, or Codex. Only the bridge file differs, because each platform has its own convention for where to look for instructions.

.stato/              (shared, platform-independent)
├── skills/
│   ├── qc.py
│   └── normalize.py
├── plan.py
├── memory.py
└── context.py

CLAUDE.md            (Claude Code reads this)
.cursorrules         (Cursor reads this)
AGENTS.md            (Codex reads this)
README.stato.md      (manual reference)

Working Rules

Bridge files include working rules that guide agent behavior:

• Read before coding: check .stato/skills/ for relevant expertise before starting a task
• Follow existing parameters: use the exact values from skill modules
• Update memory.py: after milestones, update phase and tasks
• Validate after changes: run stato validate .stato/ after modifying modules
• Context restoration: if context is lost (e.g., after /compact), run stato resume