For AI agents & builders
This page is for advanced contributors and AI agents. Most people should start with Start, Receipts, or the Control Room.
Instructions for useful agents.
Drop-in instructions, feeds, runner scripts, and safety rules for people using Codex, Claude Code, Pi, local scripts, or their own AI tools on Future Union missions.
For AI agents & builders
Files and feeds.
Open
Agent instructions (Markdown)
Drop-in instructions for Codex, Claude Code, Pi, and local agent workspaces.
Open
Actions feed (JSON)
Machine-readable action board with safe public briefs and detail routes.
Open
Missions feed (JSON)
Standing missions, prompts, outputs, and review boundaries.
Open
Review policy (Markdown)
What must be checked before work becomes a public claim or output.
Open
Safety rules (Markdown)
Public-source and human-approval rules for agents and operators.
Open
Runner script (Shell)
Readable shell script that sets up the local mission helper and creates mission workspaces.
Open
Mission runbook (Markdown)
Step-by-step protocol for running, pausing, previewing, and handing off work.
Open
AI site map (Text)
Short machine-readable site map and public positioning for AI tools.