# Meta Skill Routing > Strategies for dynamic skill discovery to minimize context overhead. Use when this capability is needed. - Skill: `tomevault-io/meta-skill-routing` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/meta-skill-routing` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/meta-skill-routing/raw - Safety review: pending (external: skill-scanner PASS, skillspector CAUTION) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Product & Planning - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/meta-skill-routing --- # Meta-Skill Routing When a library has dozens of skills, you cannot load them all. You must "route" your attention effectively. ## The Two-Step Pattern 1. **Index Lookup**: Read the root `README.md` or a dedicated index to identify which skills are relevant to the current request. 2. **Explicit Load**: Only load the content of the specific `SKILL.md` files needed. ## Routing Heuristics - **Keyword Matching**: Look for domain-specific keywords (e.g., "Flask", "TDD", "Transformers"). - **Intent Discovery**: Ask, "Is this a planning, implementing, or verifying task?" and route to the corresponding skill suite. ## Best Practices - **Atomic Loading**: Avoid "just in case" loading. Only load a skill when you are about to perform an action related to it. - **Reference Over Inclusion**: Link to skills in documentation instead of copying their content. --- > Converted and distributed by [TomeVault](https://tomevault.io/claim/jcorpac) — claim your Tome and manage your conversions.