Skills for working on the Stanford SNAP / Leskovec group's infrastructure: ILC cluster access, Stanford Office 365 email, Overleaf paper writing, and Slidev presentation authoring.
Use this skill when the user asks you to create a talk, lecture, seminar, or conference presentation from an existing outline, org-mode note, markdown document, or rough text. Produces a Slidev presentation (markdown + Vue components) with appropriate structure, animations, speaker notes, and layout choices. Do NOT use this for PowerPoint/Keynote files — this skill is Slidev-specific.
Use this skill when the user wants to read, search, send, or draft Stanford Office 365 email from the command line or via an agent. Works via the Microsoft Graph API — does NOT require Stanford IT to enable IMAP on the mailbox. Syncs to a local Maildir for use with notmuch, mutt, or similar tools.
Use this skill when the user needs to run code, submit jobs, or access data on the SNAP/ILC cluster at Stanford. Covers SLURM job submission (single `il` partition with QoS-based scheduling, correct account, GPU types), the three-tier storage system, and uv/pixi virtual environment management on compute nodes. Do NOT use this for Sherlock — ILC has a different account, storage layout, and cluster configuration.
Linus Torvalds-style code review. Brutally honest multi-layer analysis focused on data structures, good taste, cognitive load, and eliminating unnecessary abstraction. Reviews code like Linus reviews kernel patches.
Use this skill whenever the user asks you to write, expand, revise, or edit any part of a manuscript, grant proposal, fellowship application, or research statement. Defines the SNAP group's academic writing style (concise, claim-first, no padding) and the mechanics of iterating on Overleaf projects via local git sync and in-place bracketed instructions. Trigger on: "paper", "manuscript", "grant", "proposal", "fellowship", "abstract", "intro", "methods", "discussion", "reviewer response".
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Created: 2026-04-09 · Last major update: 2026-04-09
A collection of agent skills for working on the Leskovec SNAP group's infrastructure at Stanford.
AI agents (Claude Code, OpenCode, etc.) are capable of general tasks, but they lack knowledge about Stanford and SNAP-specific infrastructure and workflows. This repository is a resource for agents to provide that missing information — cluster access, email integration, writing workflows, and more.
Think of it as a shared AGENTS.md for the group: each skill teaches your agent how to do one thing the SNAP way.
This repository is packaged as a Claude Code plugin. It is both the plugin and its own single-plugin marketplace, so installation is one shell line:
claude plugin marketplace add moritzschaefer/snap-skills && \
claude plugin install snap-skills@snap-skills
After installing, run /reload-plugins inside Claude Code (or restart). Skills become available under the snap-skills: namespace, e.g. /snap-skills:ilc-access. Claude will also load them automatically when their descriptions match the task.
To update: git pull has no effect on the local cache — run /plugin marketplace update snap-skills inside Claude Code (or claude plugin marketplace update snap-skills from the shell), then /reload-plugins.
OpenCode users: OpenCode uses a flat skill layout and does not support Claude Code's plugin manifest. Clone directly and symlink each skill:
git clone [email protected]:moritzschaefer/snap-skills.git ~/code/snap-skills
for d in ~/code/snap-skills/skills/*/; do
ln -s "$d" ~/.config/opencode/skills/"$(basename $d)"
done
Note: Some skills require one-time setup (API keys, authentication, server-side config, etc.). The SKILL.md files specify these setup steps and instruct the agent to surface them to you if the skill fails to execute.
snap-skills/
├── .claude-plugin/
│ ├── plugin.json # plugin manifest
│ └── marketplace.json # single-plugin marketplace catalog
└── skills/
├── ilc-access/SKILL.md
├── email/
├── paper-writing/SKILL.md
├── create-presentation/SKILL.md
└── linus-review/SKILL.md
| Skill | Description |
|---|---|
skills/ilc-access/ | Submit SLURM jobs on the SNAP/ILC cluster with the correct account, partitions, storage tiers, and uv environment setup |
skills/email/ | Read, search, and send Stanford Office 365 email locally via Microsoft Graph API (needs testing) |
skills/paper-writing/ | Iterate on Overleaf manuscripts locally via git sync and in-place bracketed instructions; contains writing style instructions |
skills/create-presentation/ | Build Slidev presentations from an outline (org-mode note, markdown, or text) |
skills/linus-review/ | Review code changes with a Linus-inspired lens focused on data structures, special cases, cognitive load, and unnecessary abstraction |
This project lives from your updates.
Skills that are planned but not yet written. If you need one of these, open an issue — I'll prioritize it.
npx claudepluginhub snap-stanford/snap-skills --plugin snap-skillsComprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Design fluency for frontend development. 1 skill with 23 commands (/impeccable polish, /impeccable audit, /impeccable critique, etc.) and curated anti-pattern detection.
Behavioral guidelines to reduce common LLM coding mistakes, derived from Andrej Karpathy's observations on LLM coding pitfalls
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.