# Orchestration > Coordinates other skills and agents. ALWAYS use this skill on startup. Use when this capability is needed. - Skill: `tomevault-io/orchestration-3` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/orchestration-3` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/orchestration-3/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/orchestration-3 --- # Skill Guidelines Invoke these skills liberally - most tasks use multiple skills: | Skill | When to use | | ----------------------------------- | ------------------------------------- | | `@code:code-quality` | Before editing code | | `@code:cli` | When running shell commands | | `@build:precommit` | Before running builds or tests | | `@git:git-workflow` | For all git operations | | `@orchestration:conversation-style` | For response guidelines | | `@orchestration:llm-context` | When working with `.llm/` directories | ## Git Commits **ALWAYS** delegate to the `@git:commit-handler` agent for all commit operations. Never run `git commit` directly. **ALWAYS** delegate to the `@git:conflict-resolver` agent to resolve any git merge or rebase conflicts. **ALWAYS** delegate to the `@git:rebaser` agent to rebase the current branch on upstream. ## File Writing Policy **NEVER** write files to `/tmp` or other system temporary directories - reading from `/tmp` triggers permission prompts. Write scratch files and temporary outputs to `.llm/` instead. ## Workflow Orchestration Always run `/orchestration:finish` before returning control to the user. It handles the case where there's nothing to do. --- > Converted and distributed by [TomeVault](https://tomevault.io/claim/motlin) — claim your Tome and manage your conversions.