trio-bridge
01Command Claude Code
Scan the current project in the background and write deduplicated Trio proposal cards into proposals/.
Research-hardened two-agent coding loop (Lead + adversarial Evaluator) for Claude Code, with a portable driver for Codex, Cursor, OpenCode, Pi, Hermes, Gemini and more — agents communicate via markdown mailboxes
This repository also configures its own agents. See what trio-agent-loop tells them →
Command Claude Code
Scan the current project in the background and write deduplicated Trio proposal cards into proposals/.
Command Claude Code
Initialize the Trio loop in the current project — creates the loop/ mailbox directory and GOAL.md from the given goal text.
Command Claude Code
Run the production-readiness graph through the shared Trio productionize procedure.
Command Claude Code
Manually recover an orphaned SHIP verdict — performs the Evaluator's retirement commit (product + mailbox) for a SHIP whose commit: lines are missing. No subagent, no model config.
Command Claude Code
Run one or more iterations of the Trio agent loop using the loop/ mailbox files. /trio = one supervised iteration; /trio auto = keep iterating in-session until SHIP, BLOCKED, or NEEDSHUMAN.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: