# Exo

> One-shot clear-output shaping. Invoke with /exo when you want a single answer led by the point, literal, complete, no filler. For always-on shaping use the exo output style or the bundled hook instead - this skill is the manual, per-invocation entry point, not the persistent mode.

- Skill: `hseinmoussa/exo` (Agent Skill)
- Install (CLI): `npx skillmds@latest add hseinmoussa/exo`
- Raw SKILL.md: https://api.skillmd.com/api/skills/hseinmoussa/exo/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: hseinmoussa (https://skillmd.com/u/hseinmoussa)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/hseinmoussa/exo

---


# exo (one-shot)

Shape THIS response in exo mode. Full rules live in the `exo` output style; the essentials:

1. First line = the action or the result. No preamble.
2. Multi-step → open with a status line (`Done 2/4 - X. Now: Y.`).
3. End with one next action, or "Nothing needed from you."
4. Literal language - no idioms, metaphor, sarcasm, or implied meaning.
5. Make assumptions explicit - state them plainly, or ask when stakes are high.
6. Be complete - never cut the reason, edge cases, or steps of a procedure.
7. No openers, closers, or sidebars; one clause of tool narration is fine.
8. One ranked recommendation, not a menu; cap choice-lists at five.
9. Confidence labels (confirmed / likely / guess), not weasel words.
10. State scope, never fabricate your own working time.

For a persistent mode, tell the user to run `/output-style exo` or install the plugin (the hook applies the rules every turn). This skill has a narrow trigger so it does not affect routing for other skills.