Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Write professional-grade skills for agents, validate them using LLMs, and maintain a lean context window.
| Date | Stars |
|---|---|
| 2026-08-01 | 2166 |
| 2026-08-06 | 2166 |
| 2026-08-11 | 2207 |
| 2026-08-18 | 2226 |
| 2026-08-19 | 2227 |
| 2026-08-20 | 2229 |
| 2026-08-22 | 2231 |
| 2026-08-23 | 2232 |
| 2026-08-24 | 2234 |
| 2026-08-25 | 2235 |
| 2026-08-26 | 2236 |
| 2026-08-27 | 2237 |
| 2026-08-28 | 2238 |
| 2026-08-29 | 2241 |
| 2026-08-31 | 2242 |
| 2026-09-01 | 2244 |
| 2026-09-03 | 2246 |
| 2026-09-04 | 2248 |
| 2026-09-05 | 2246 |
| 2026-09-06 | 2247 |
| 2026-09-07 | 2248 |
| 2026-09-08 | 2249 |
| 2026-09-09 | 2251 |
| 2026-09-10 | 2252 |
| 2026-09-12 | 2253 |
| 2026-09-14 | 2254 |
| 2026-09-16 | 2254 |
| 2026-09-17 | 2255 |
| 2026-09-18 | 2254 |
| 2026-09-20 | 2254 |
| 2026-09-21 | 2255 |
| 2026-09-22 | 2257 |
| 2026-09-23 | 2262 |
| 2026-09-24 | 2263 |
| 2026-09-28 | 2265 |
| 2026-09-29 | 2267 |
| 2026-09-30 | 2268 |
| 2026-10-04 | 2269 |
| 2026-10-06 | 2268 |
| 2026-10-07 | 2270 |
| 2026-10-09 | 2273 |
| 2026-10-10 | 2271 |
| 2026-10-11 | 2272 |
| 2026-10-12 | 2273 |
| 2026-10-14 | 2274 |
| 2026-10-15 | 2275 |
| 2026-10-16 | 2276 |
| 2026-10-17 | 2276 |
| 2026-10-20 | 2278 |
| 2026-10-21 | 2278 |
| 2026-10-23 | 2279 |
| 2026-10-24 | 2278 |
| 2026-10-27 | 2279 |
| 2026-10-28 | 2281 |
| 2026-10-29 | 2283 |
| 2026-11-01 | 2284 |
| 2026-11-02 | 2284 |
| 2026-11-03 | 2285 |
| 2026-11-04 | 2287 |
Today
+2 stars today
This week
+6 stars this week
This month
+18 stars this month
Momentum
0.0
growth rate 0.26%/day
# Best Practices for Creating Agent Skills This guide explains how to write professional-grade skills for agents, validate them using LLMs, and maintain a lean context window. This guide is a concentrated set of best practices for creating agent skills. If you're looking for a comprehensive documentation see [Claude's docs](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/best-practices). **To evaluate if your skills do well and prevent regressions, check out [skillgrade](https://github.com/mgechev/skillgrade).** ## Structure of a skill Every skill must follow this directory structure: Plaintext ``` skill-name/ ├── SKILL.md # Required: Metadata + core instructions (<500 lines) ├── scripts/ # Executable code (Python/Bash) designed as tiny CLIs ├── references/ # Supplementary context (schemas, cheatsheets) └── assets/ # Templates or static files used in output ``` * **SKILL.md:** Acts as the "brain." Use it for navigation and high-level procedures. * **References:** Link directly from SKILL.md. Keep them **one level deep** only. * **Scripts:** Use for fragile/repetitive operations where variation is a bug. **Do not bundle library code here**; ## Optimize the frontmatter for discoverability The `name` and `description` in the frontmatter of your `SKILL.md` are the only fields that the agent sees before triggering a skill. If they are not optimized for discoverability and specific enough, your skill is invisible. * **Adhere to Strict Naming:** The name field must be 1-64 characters, contain only lowercase letters, numbers, and hyphens (no consecutive hyphens), and **must exactly match the parent directory name** (e.g., name: `angular-testing` must live in `angular-testing/SKILL.md`). * **Write Trigger-Optimized Descriptions:** (Max 1,024 characters). This is the only metadata the agent sees for routing. Describe the capability in the third person and include "negative triggers." * **Bad:** "React skills." (Too vague). * **Good:** "Creates and builds React components using Tailwind CSS. Use when the user wants to update component styles or UI logic. Don't use it for Vue, Svelte, or vanilla CSS projects." ## Progressive disclosure and resource management Maintain a pristine context window by loading information only when needed. **SKILL.md** is the "brain" for high-level logic; offload details to subdirectories. * **Keep SKILL.md Lean:** Limit the main file to **\<500 lines**. Use it for navigation and primary procedures. * **Use Flat Subdirectories:** Move bulky context to standard folders. Keep files exactly **one level deep** (e.g., `references/schema.md`, not `references/db/v1/schema.md`). * `references/`: API docs, cheatsheets, domain logic. * `scripts/`: Executable code for deterministic tasks. * `assets/`: Output templates, JSON schemas, images. * **Just-in-Time (JiT) Loading:** Explicitly instruct the agent when to read a file. It will not see these resources until you direct it to (e.g., *"See `references/auth-flow.md` for specific error codes"*). * **Explicit Pathing:** Always use **relative paths** with forward slashes (`/`), regardless of the OS. Skills are for agents, not humans. To keep the context window lean and avoid unnecessary token consumption. **Do not create:** * **Documentation files:** `README.md`, `CHANGELOG.md`, or `INSTALLATION_GUIDE.md`. * **Redundant logic:** If the agent already handles a task reliably without help, delete the instruction. * **Library code:** Skills should reference existing tools or contain tiny, single-purpose scripts. Long-lived library code belongs in standard repo CLI directories. ## Use specific procedural instructions instead of prose Create instructions for LLMs instead of humans. * **Use Step-by-Step Numbering:** Define the workflow as a strict chronological sequence. If there is a decision tree, map it out clearly (e.g., *"Step 2: If you need source maps run
Excerpt of 10,474 characters
Read on GitHub
Minko Gechev · Google · United States
8
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:30e0c87a3afbbfc6, llm:Repository description: 'Write professional-grade skills for agents, validate them using LLMs, and maintain a lean context window.' Language: Python. No topics provided.
matched fp:30e0c87a3afbbfc6, llm:Repository description: 'Write professional-grade skills for agents, validate them using LLMs, and maintain a lean context window.' Language: Python. No topics provided.
matched fp:30e0c87a3afbbfc6, llm:Repository description: 'Write professional-grade skills for agents, validate them using LLMs, and maintain a lean context window.' Language: Python. No topics provided.