Initialize a Codex Trio mailbox from the user's goal. Use when the user explicitly asks to initialize Trio or when the Trio skill needs a new mailbox.
Scanned 9/5/2026
Install to Claude Code
npx -y skills add albiol2004/trio-agent-loop --skill trio-init --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Trio Init?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/albiol2004-trio-init-275ee88c)More formats (shields.io, HTML) on the badges page.
---
name: trio-init
description: Initialize a Codex Trio mailbox from the user's goal. Use when the user explicitly asks to initialize Trio or when the Trio skill needs a new mailbox.
---
# Initialize a Codex Trio mailbox
Create the mailbox in `loop/`, or in the requested `loop-<name>/` directory.
Never overwrite or repurpose a mailbox whose STATE says it may be running.
Write:
- `GOAL.md`: `profile: software|data`, mission, objective definition of done,
and constraints. Data goals must name ground truth, reconciliation tolerance,
key uniqueness, and rerun/idempotence expectations.
- `STATE.md`: `schema: 1`, iteration 0, max_iterations 10, ready status, mission fingerprint,
rejected approaches, and key decisions.
- `LOG.md`: `# Trio loop log`.
- Empty `PLAN.md`, `REPORT.md`, and `VERDICT.md`.
Ask whether the mailbox should be committed or ignored. Then tell the user
they can simply say `Run a trio loop on <goal>`; they do not need to invoke
this skill separately in normal use.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!
Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...