Hermes Registry

hermes-skill-factory

v1.0.0Skill

Use when you want to auto-generate reusable skills from repeated workflows. Watches your Hermes sessions, detects patterns, and proposes SKILL.md + plugin.py pairs on demand.

IDhermes-skill-factory

Hermes Skill Factory

Overview

Skill Factory is a meta-skill that watches your Hermes sessions and automatically turns repeated workflows into reusable skills. Every time you solve a problem with Hermes — setting up a project, debugging code, creating a PR — that workflow normally disappears at session end. Skill Factory detects patterns across your work and, at the right moment, proposes a structured skill for you to accept or skip.

Built for Hermes Agent v2026.3+.

When to Use

  • You find yourself explaining the same workflow to Hermes more than once
  • You want to capture a workflow you just completed as a reusable skill
  • You want to generate both a SKILL.md (AI instructions) and a plugin.py (slash command) in one step
  • Don't use for: one-off tasks unlikely to repeat, or skills that already exist in the registry

Installation

git clone https://github.com/Romanescu11/hermes-skill-factory
cd hermes-skill-factory
bash install.sh

Or manually:

mkdir -p ~/.hermes/skills/meta/skill-factory
cp skills/skill-factory/SKILL.md ~/.hermes/skills/meta/skill-factory/
cp plugins/skill_factory.py ~/.hermes/plugins/

Then activate:

hermes skills reload
hermes skills enable skill-factory

Usage

Skill Factory runs silently in the background during every session. When it detects a repeatable pattern it prompts:

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🏭 SKILL FACTORY — New Skill Detected ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
I noticed you repeatedly set up a Python environment, installed dependencies, and ran tests in the same order.

Proposed Skill: python-env-setup
Category:       software-development
Description:    Reproducible Python project setup workflow

What it captures:
  1. Create venv and activate
  2. Upgrade pip and install dependencies
  3. Run pytest to verify environment

Generate: [A] SKILL.md  [B] plugin.py  [C] Both  [D] Skip
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Choose C and it writes both files immediately:

  • ~/.hermes/skills/<category>/<name>/SKILL.md — AI instructions for the workflow
  • ~/.hermes/plugins/<name>.py — a slash command that triggers it directly

Slash Commands

CommandDescription
/skill-factory proposeManually trigger a proposal from observed patterns
/skill-factory listList all proposed skills in the queue
/skill-factory statusShow observer status and session stats
/skill-factory queueView pending proposals
/skill-factory save <name>Save a specific named proposal
/skill-factory clearClear the proposal queue

You can also use natural language:

  • "Save this as a skill"
  • "Remember how to do this"
  • "Turn this workflow into a reusable skill"

What Gets Generated

SKILL.md

A complete skill definition following Hermes' native skill format:

---
name: Python Env Setup
category: software-development
description: Reproducible Python project setup
tags: [python, venv, testing]
---
# Python Env Setup
## When to Activate
...
## Workflow
### Phase 1: Environment
1. python -m venv .venv
2. source .venv/bin/activate
...
## Examples
...

plugin.py

A scaffolded Hermes plugin with a slash command:

def register(hermes):
    @hermes.command(name="python-env-setup", ...)
    async def run_skill(ctx, args=""):
        # Step 1: Create venv
        # Step 2: Install deps
        # Step 3: Run tests
        ...

Repo Structure

hermes-skill-factory/
├── skills/
│   └── skill-factory/
│       └── SKILL.md          # The meta-skill (core AI instructions)
├── plugins/
│   └── skill_factory.py      # Plugin: /skill-factory commands
├── templates/
│   ├── SKILL_TEMPLATE.md     # Template for generated skills
│   └── PLUGIN_TEMPLATE.py    # Template for generated plugins
├── examples/
│   └── generated/
│       └── git-pr-workflow/  # Example Skill Factory output
│           └── SKILL.md
├── docs/
│   └── how-it-works.md
└── install.sh

Common Pitfalls

  1. Wrong Hermes version. Requires v2026.3+. Check with hermes --version before installing.
  2. Skill not activating. Run hermes skills reload and hermes skills enable skill-factory after install.
  3. Generated skill path. Skills are written to ~/.hermes/skills/, not the in-repo skills/ tree. Use hermes skills publish if you want to share to the registry.
  4. Plugin not available. Ensure skill_factory.py is in ~/.hermes/plugins/ and restart the session.
  5. Duplicate proposals. If Skill Factory proposes a skill that already exists, use /skill-factory clear to discard and add a note to your session so it doesn't re-propose.

Verification Checklist

  • Hermes Agent v2026.3+ installed (hermes --version)
  • install.sh ran without errors, or files copied manually
  • hermes skills reload and hermes skills enable skill-factory executed
  • /skill-factory status returns active observer
  • A test proposal can be triggered via /skill-factory propose
  • Generated SKILL.md passes frontmatter validation (name, description present, description ≤ 1024 chars)