The official seed skill library for EurekaClaw

🌱 What are seed skills?

Seed skills is a curated collection of foundational skills for the EurekaClaw AI agent ecosystem. Each skill is a self-contained directory with a SKILL.md file that teaches the agent a new capability — from rigorous mathematical proof techniques to structured literature decomposition.

---

📦 Skill Catalog

Skill	Description
compactness_argument	Techniques for constructing compactness-based proofs
complexity_template	Structured templates for complexity analysis
concentration_inequalities	Applying Hoeffding, Bernstein, and related bounds
contradiction_proof	Proof by contradiction strategies and patterns
eluder_dimension	Eluder dimension reasoning for function classes
empirical_validation	Empirical testing and experimental validation workflows
fundamental_inequalities	Core inequalities: Cauchy-Schwarz, Jensen, AM-GM, etc.
hypothesis_gen	Systematic hypothesis generation for research
induction_strategy	Mathematical induction — base cases, inductive steps, strong induction
latex_algorithm_and_syntax	Writing algorithms and proofs in clean LaTeX
literature_decomp	Decomposing and synthesizing research papers
nonlinear_dueling_bandits_general_function_approximation	Dueling bandits with general nonlinear function classes
overdamped_langevin_convergence	Convergence analysis for overdamped Langevin dynamics
paper_structure	Academic paper structure and writing discipline
regularized_bandit_fast_rates	Fast convergence rates in regularized bandit settings
rigorous_proof_citation_discipline	Citation standards and proof rigor in academic writing
underdamped_langevin_convergence	Convergence analysis for underdamped (kinetic) Langevin

---

🚀 Quick Start

Install the skills to your EurekaClaw

The seed skills will be automatically install at EurekaClaw onboarding. To update the seed skills, run

eurekaclaw install-skills --force

🏗️ Skill Structure

Every skill follows the OpenClaw-compatible format:

my_skill/
└── SKILL.md        # Required: Instructions, tools, and workflow for the agent

SKILL.md contains YAML frontmatter (name, description, requirements) followed by markdown instructions that the agent reads at runtime.

---

🤝 Contributing

We welcome new skills! Two ways to contribute:

Option 1 — Pull Request (this repo)

# 1. Fork and clone
git clone https://github.com/EurekaClaw/seed_skills.git

# 2. Create your skill directory
mkdir seedskills/my_new_skill
touch seedskills/my_new_skill/SKILL.md   # Write your skill here

# 3. Submit a PR
git checkout -b add-my-new-skill
git add seedskills/my_new_skill/
git commit -m "feat: add my_new_skill"
git push origin add-my-new-skill

Option 2 — Publish directly to ClawHub

Upload your skill to the public registry and share it with the entire OpenClaw community.

Skill quality bar

A good seed skill should:

• ✅ Have a clear, single responsibility
• ✅ Include concrete examples or worked demonstrations in SKILL.md
• ✅ Be general enough to be reusable across research contexts
• ✅ Cite sources or foundational references where relevant

---

🧪 Testing

This repo enforces that every skill directory contains a valid SKILL.md via automated CI:

# Run locally
pip install pytest
pytest tests/ -v

CI runs on every push and pull request via GitHub Actions. A PR that adds a skill folder without a SKILL.md will fail automatically.

---

📚 Learn More

Resource	Link
EurekaClaw main repo	github.com/EurekaClaw

---

⚖️ License

Apache 2.0 — see LICENSE for details.

---

_{Built with 🦞 by EurekaClaw}