🧩 Awesome Hermes Skills — Core Four

Four production-ready skills for Hermes Agent — no fluff, no extras. Each one earns its place.

---

🪽 Artifact Preview · artifact-preview/

Write code. See it live. Instantly. Never lose a version. 🔥

The “Claude Artifacts” experience — for any AI agent.

Write a prompt → get a polished, interactive HTML artifact → Chrome opens automatically with live reload. Portrait, horizontal, or full-screen modes. Auto-detects the right layout from your HTML. Inline editor. Retina screenshots. Zero config.

One-line install:

curl -fsSL https://raw.githubusercontent.com/ChuckSRQ/awesome-hermes-skills/v4.3/artifact-preview/install.sh | bash

Uninstall (also one-line):

curl -fsSL https://raw.githubusercontent.com/ChuckSRQ/awesome-hermes-skills/v4.3/artifact-preview/uninstall.sh | bash

---

🔬 Deep Research · deep-research/

9 lenses. One question. Analysis that single Google searches can’t touch.

A structured multi-angle research engine. Forces you to think about any question from 9 fundamentally different angles — each lens surfaces things the others miss.

Output: executive summary · deep-dive · key players · open questions

Perfect for: investment research, competitive analysis, architecture decisions, market sizing, due diligence, understanding emerging tech

Invocation: do deep research on [your question]

Knowledge base: ~/research-skill-graph/

---

🧪 Agentic Benchmark Testing · agentic-self-improvement/

Closed-loop behavioral benchmarking + evolutionary self-improvement

Run behavioral benchmarks against the agent. Detect failures. Generate patches. Apply with auto-revert safety net.

Implements the full GEPA (Genetic-Evolutionary Prompt Adjustment) optimization loop with DSPy integration. Turns real conversation failures into automated tests, then evolves skills end-to-end.

Usage:

/self-improve                              # run all benchmarks, show patches
/self-improve --mode=apply                 # apply + auto-revert if regression
/self-improve --benchmarks=mandatory_tool  # specific benchmark
/self-improve --revert                     # revert last patch

---

⚡ GH Copilot · gh-copilot/

GitHub Copilot CLI plugin — 23 Hermes development skills.

A plugin for gh copilot CLI that wires 23 production skills directly into GitHub Copilot.

Invoke any skill via /skill or gh copilot -- -i "use ".

GH PR Workflow (key rules)

gh copilot -- -i "use github-pr-workflow to open a PR for this fix" --allow-all
gh copilot -- -i "use code-review to review the current branch" --allow-all

---

📁 Repo Structure

awesome-hermes-skills/
├── artifact-preview/           # Live HTML preview server + Chrome launcher
├── deep-research/              # 9-lens research engine
├── agentic-self-improvement/   # Behavioral benchmarking + GEPA optimization
├── gh-copilot/                # GitHub Copilot CLI plugin (23 skills)
├── README.md                   # This file
├── LICENSE                     # MIT
└── agents.md                   # Repo operating rules

---

⚙️ Setup

# Clone
git clone https://github.com/ChuckSRQ/awesome-hermes-skills.git ~/.hermes/skills

# Individual skills
cp -r awesome-hermes-skills/artifact-preview ~/.hermes/skills/
cp -r awesome-hermes-skills/deep-research ~/.hermes/skills/
cp -r awesome-hermes-skills/agentic-self-improvement ~/.hermes/skills/
cp -r awesome-hermes-skills/gh-copilot ~/.hermes/skills/

---

📜 License

MIT — use freely, modify freely, credit this repo if you’re feeling generous.

Maintained with ❤️ for the Hermes Agent community GitHub · Hermes Agent