docs: add Hermes install guidance

Author: GarethManning

.agents/plugins/marketplace.json

@@ -14,7 +14,7 @@
       "category": "Education",
       "interface": {
         "displayName": "Education Agent Skills",
-        "shortDescription": "131 evidence-based skills for curriculum design, lesson planning, and assessment"
+        "shortDescription": "165 evidence-based skills for curriculum design, lesson planning, and assessment"
       }
     }
   ]

.claude-plugin/marketplace.json

@@ -5,7 +5,7 @@
     "url": "https://github.com/GarethManning"
   },
   "metadata": {
-    "description": "131 evidence-based pedagogical skills for educators. Curriculum design, assessment, learning science, wellbeing, professional learning, historical thinking, and more.",
+    "description": "165 evidence-based pedagogical skills for educators. Curriculum design, assessment, learning science, wellbeing, professional learning, historical thinking, and more.",
     "version": "2.1.0"
   },
   "plugins": [

.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "education-agent-skills",
-  "display_name": "Claude Education Skills Library",
+  "display_name": "Education Agent Skills Library",
   "version": "2.1.0",
-  "description": "131 evidence-based pedagogical skills for educators. Curriculum design, assessment, learning science, wellbeing, professional learning, historical thinking, and more.",
+  "description": "165 evidence-based pedagogical skills for educators. Curriculum design, assessment, learning science, wellbeing, professional learning, historical thinking, and more.",
   "author": {
     "name": "Gareth Manning",
     "url": "https://github.com/GarethManning"

.codex-plugin/plugin.json

@@ -1,13 +1,13 @@
 {
   "name": "education-agent-skills",
   "version": "2.1.0",
-  "description": "131 evidence-based education skills for curriculum design, lesson planning, assessment, and AI-assisted teaching — built on the Agent Skills open standard.",
+  "description": "165 evidence-based education skills for curriculum design, lesson planning, assessment, and AI-assisted teaching — built on the Agent Skills open standard.",
   "repository": "https://github.com/GarethManning/education-agent-skills",
   "license": "CC-BY-SA-4.0",
   "skills": "./skills/",
   "interface": {
     "displayName": "Education Agent Skills",
-    "shortDescription": "131 evidence-based skills for curriculum design, lesson planning, and assessment",
+    "shortDescription": "165 evidence-based skills for curriculum design, lesson planning, and assessment",
     "composerIcon": "./assets/education-agent-skills-icon.svg"
   }
 }

AGENTS.md

@@ -1,6 +1,6 @@
 # AGENTS.md — Guidance for coding agents working on this repo
 
-This repo is a library of 131 evidence-based education skills across 17 domains, packaged for use in Claude Code, Claude.ai (via MCP), OpenAI Codex, and any Agent Skills-compatible tool.
+This repo is a library of 165 evidence-based education skills across 20 domains, packaged for use in Claude Code, Claude.ai (via MCP), OpenAI Codex, Hermes Agent, and any Agent Skills-compatible tool.
 
 ## Skill structure
 

CLAUDE.md

@@ -1,7 +1,7 @@
 # Education Agent Skills Library — Claude Code Entry Point
 
 ## What this is
-152 evidence-based education skills across 19 domains for curriculum design, lesson planning, and assessment. CC BY-SA 4.0. Works in Claude Code, Claude.ai (via MCP), and OpenAI Codex. Includes a live MCP server at mcp-server-sigma-sooty.vercel.app/mcp, a CoWork plugin (.claude-plugin/), and a Codex plugin (.codex-plugin/).
+165 evidence-based education skills across 20 domains for curriculum design, lesson planning, assessment, and student-facing AI learning support. CC BY-SA 4.0. Works in Claude Code, Claude.ai (via MCP), OpenAI Codex, and Hermes Agent. Includes a live MCP server at mcp-server-sigma-sooty.vercel.app/mcp, a CoWork plugin (.claude-plugin/), a Codex plugin (.codex-plugin/), and Hermes selected-install guidance.
 
 ## Stack
 TypeScript (77.9%), Python (16.7%). Playwright for testing. Skills are structured markdown (SKILL.md) with YAML frontmatter. MCP server in mcp-server/ subdirectory. CI via GitHub Actions.
@@ -24,15 +24,15 @@ After adding or editing ANY skill, these three steps are mandatory before commit
 3. Stage all three files: the SKILL.md, registry.json, AND mcp-server/src/skills.json
 CI will fail if the bundle is out of sync with the skills. The MCP server serves a pre-built snapshot — unbundled changes will NOT appear on the live server even after Vercel redeploy.
 
-## 19 domains
-ai-learning-science, ai-literacy, curriculum-alignment, curriculum-assessment, eal-language-development, environmental-experiential-learning, explicit-instruction, global-cross-cultural-pedagogies, historical-thinking, inclusive-design, literacy-critical-thinking, memory-learning-science, montessori-alternative-approaches, original-frameworks, professional-learning, questioning-discussion, self-regulated-learning, systems-thinking, wellbeing-motivation-agency
+## 20 domains
+ai-learning-science, ai-literacy, curriculum-alignment, curriculum-assessment, eal-language-development, environmental-experiential-learning, explicit-instruction, global-cross-cultural-pedagogies, historical-thinking, inclusive-design, literacy-critical-thinking, memory-learning-science, montessori-alternative-approaches, original-frameworks, professional-learning, questioning-discussion, self-regulated-learning, student-learning, systems-thinking, wellbeing-motivation-agency
 
 ## Key conventions
 - Every skill must cite named research — no skills without evidence grounding
 - Evidence strength rated: strong, moderate, emerging, or original (honestly labelled)
 - YAML schema headers mandatory: skill_id, domain, evidence_strength, evidence_sources, input_schema, output_schema, chains_well_with
 - Skills live in skills/<domain>/<skill-name>/SKILL.md
-- Original frameworks (Domain 13) clearly labelled as practitioner-originated
+- Original frameworks (Domain 15) clearly labelled as practitioner-originated
 - Orchestrator/composite skills always have: Evidence Space, Component Evidence, Synthesis Evidence, Appropriate Use, and Dependency Maintenance sections. The composite pathway is always labelled emerging until independently tested.
 
 ## Commit

CONTRIBUTING.md

@@ -0,0 +1,84 @@
+# Contributing
+
+Thank you for considering a contribution to the Education Agent Skills Library. The bar is intentionally high: each skill should be evidence-grounded, practically useful, and clear enough for both educators and AI agents to use safely.
+
+## What belongs here
+
+Good contributions usually add or improve a reusable education skill that supports one of these needs:
+
+- curriculum design;
+- assessment design and validity;
+- learning science and memory;
+- self-regulated learning and metacognition;
+- literacy, questioning, dialogue, and critical thinking;
+- wellbeing, motivation, belonging, and student agency;
+- inclusive design;
+- AI literacy or AI learning support;
+- student-facing study interactions;
+- professional learning;
+- systems thinking or original practitioner frameworks that are clearly labelled.
+
+## Inclusion criteria
+
+A skill should:
+
+1. solve a recurring education design or learning-support problem;
+2. be useful beyond one school, course, or lesson;
+3. include named evidence sources where it claims to be research-grounded;
+4. label emerging, mixed, practitioner, or original-framework evidence honestly;
+5. preserve educator judgement rather than pretending the skill can decide everything;
+6. include clear input and output schemas;
+7. include practical quality checks and common pitfalls;
+8. avoid unsupported frameworks, especially learning-styles/VAK-style claims.
+
+See [`docs/EXCLUSIONS.md`](docs/EXCLUSIONS.md) for excluded ideas and [`docs/EVIDENCE.md`](docs/EVIDENCE.md) for evidence standards.
+
+## Skill structure
+
+Skills live at:
+
+```text
+skills/<domain>/<skill-name>/SKILL.md
+```
+
+Do not rename or move existing skill folders unless the change is part of a deliberate repository-wide migration.
+
+Every `SKILL.md` must preserve:
+
+- YAML frontmatter;
+- `name` and `description` fields required by the Agent Skills standard;
+- evidence-strength metadata;
+- input and output schemas;
+- any existing citations and limitations.
+
+## Validation before a pull request
+
+After adding or editing skills, run:
+
+```bash
+python3 scripts/generate-registry.py
+python3 scripts/validate-registry.py
+python3 scripts/validate-skills.py
+cd mcp-server && npm run bundle-skills && npm run build && npm test
+cd ..
+npm test
+```
+
+Commit generated files when they change:
+
+```text
+registry.json
+mcp-server/src/skills.json
+mcp-server/dist/skills.json
+```
+
+## Pull request guidance
+
+In your PR, briefly explain:
+
+- what skill or documentation changed;
+- why the change is educationally useful;
+- what evidence or practitioner framework it relies on;
+- what validation commands you ran.
+
+Small, well-evidenced improvements are better than broad additions that dilute the library. The point is not to have the most skills; it is to have skills worth trusting.

README.md

@@ -5,7 +5,7 @@
 [![License: CC BY-SA 4.0](https://img.shields.io/badge/License-CC%20BY--SA%204.0-lightgrey.svg)](https://creativecommons.org/licenses/by-sa/4.0/)
 [![Last Commit](https://img.shields.io/github/last-commit/GarethManning/education-agent-skills)](https://github.com/GarethManning/education-agent-skills/commits/main)
 
-An open-source library of 165 evidence-based pedagogical skills across 20 domains — works in Claude Code, Claude.ai (via MCP), and OpenAI Codex, and engineered for AI agent orchestration. Domains 1–19 are teacher and designer-facing. Domain 20 is the first student-facing domain: live AI interaction patterns that shape how AI responds to learners during study sessions.
+An open-source library of 165 evidence-based pedagogical skills across 20 domains — works in Claude Code, Claude.ai (via MCP), OpenAI Codex, and Hermes Agent, and is engineered for AI agent orchestration. Domains 1–19 are teacher and designer-facing. Domain 20 is the first student-facing domain: live AI interaction patterns that shape how AI responds to learners during study sessions.
 
 > [!IMPORTANT]
 > **Hosted MCP access now requires an auth token.**
@@ -18,7 +18,7 @@ An open-source library of 165 evidence-based pedagogical skills across 20 domain
 
 ## Get Started
 
-Works with Claude, Codex, and any tool that supports the Agent Skills standard.
+Works with Claude, Codex, Hermes Agent, and any tool that supports the Agent Skills standard.
 
 For sustainable free use, install or copy the skills locally from GitHub where possible. The hosted MCP server is a convenience endpoint for remote clients, not a requirement for using the library.
 
@@ -71,6 +71,23 @@ cp -r skills/memory-learning-science/spaced-practice-scheduler ~/.codex/skills/
 
 Full Codex guide: [docs/CODEX.md](docs/CODEX.md).
 
+### Hermes Agent
+
+Hermes users should keep this repository as the canonical source, then install only the skills they actually need locally.
+
+```bash
+hermes skills tap add GarethManning/education-agent-skills
+hermes skills install \
+  GarethManning/education-agent-skills/skills/original-frameworks/learning-target-authoring-guide \
+  --category education --yes
+```
+
+Recommended starting point: install selected skills or a small starter pack rather than all 165 skills. That keeps your local Hermes index useful instead of noisy.
+
+If you want intelligent skill discovery and recommendation rather than local/offline installs, use the hosted MCP server's `find_skills` and `suggest_skills` tools. The MCP route and the Hermes tap serve different adopter types; a separate Hermes plugin is not currently planned.
+
+Full Hermes guide: [docs/HERMES.md](docs/HERMES.md).
+
 ### Any Agent Skills-compatible tool
 
 Copy skill folders from `skills/` into your agent's skills directory. Each skill is a folder containing `SKILL.md` with name/description frontmatter — no dependencies, no build step.
@@ -95,7 +112,7 @@ I'd love to hear your thoughts. If you have suggestions, find bugs, or want to c
 ---
 
 **I'm an educator — [start here](#try-it-now)**
-No setup required. Install the plugin and start teaching.
+No setup required. Use the plugin, a local skill install, or manual copy-paste and start teaching.
 
 **I'm a developer or AI builder — [start here](#architecture)**
 YAML schemas, typed inputs and outputs, chaining metadata, [live MCP server](#mcp-server).
@@ -132,9 +149,9 @@ That is one use case. The same library can power school-wide curriculum audits,
 
 ## Try It Now
 
-### With the plugin (recommended)
+### With a runtime install (recommended)
 
-Install the plugin, then tell Claude what you need in plain language. The skills activate automatically.
+Install the skills in Claude, Codex, or Hermes, then tell your agent what you need in plain language. The relevant skills can be selected automatically or searched locally.
 
 **Example:** Say *"I'm planning a Year 9 science unit on cells — 6 weeks, 3 lessons a week."*
 

docs/AGENT_SKILLS.md

@@ -1,6 +1,6 @@
 # Agent Skills Developer Guide
 
-This document is for developers and AI builders who want to consume, orchestrate, or extend the Claude Education Skills Library programmatically.
+This document is for developers and AI builders who want to consume, orchestrate, or extend the Education Agent Skills Library programmatically.
 
 ---
 
@@ -12,7 +12,7 @@ This document is for developers and AI builders who want to consume, orchestrate
 /plugin install GarethManning/education-agent-skills
 ```
 
-This makes all 131 skills available in your Claude Code session. Skills with `disable-model-invocation: false` will auto-activate when conversation context matches their trigger descriptions. Skills with `disable-model-invocation: true` must be invoked explicitly.
+This makes all 165 skills available in your Claude Code session. Skills with `disable-model-invocation: false` will auto-activate when conversation context matches their trigger descriptions. Skills with `disable-model-invocation: true` must be invoked explicitly.
 
 ### From Source
 
@@ -35,7 +35,7 @@ The registry is the primary programmatic interface to the library. It is generat
   "version": "2.0",
   "generated": "<ISO timestamp>",
   "standard": "agent-skills-1.0",
-  "total_skills": 131,
+  "total_skills": 165,
   "domains": [
     { "id": "memory-learning-science", "label": "Memory & Learning Science", "skill_count": 8 }
   ],

docs/ARCHITECTURE.md

@@ -25,7 +25,7 @@ The YAML schemas are not documentation. They are the interface definition. Any s
 
 ### Layer 1 — The Skill Library (this repository)
 
-131 skills across 17 domains. Each skill encodes a specific, evidence-grounded instructional or curriculum design decision — retrieval practice scheduling, rubric construction, scaffolded task modification, adaptive hint design, and more.
+165 skills across 20 domains. Each skill encodes a specific, evidence-grounded instructional or curriculum design decision — retrieval practice scheduling, rubric construction, scaffolded task modification, adaptive hint design, student-facing study support, and more.
 
 An educator uses a skill by providing context through the input fields. Claude returns a structured output ready to use: a spaced practice schedule, a set of retrieval questions, a rubric with co-construction plan, a scaffolded task adapted for a specific language proficiency level.
 
@@ -67,7 +67,7 @@ The `chains_well_with` field in every skill's YAML header provides explicit chai
 
 ## MCP Server
 
-An MCP server exposes the full skill library as callable tools and prompts. Any MCP-compatible client can discover all 131 skills, read their input requirements, call them with structured parameters, and receive typed outputs — removing the manual copy-paste step entirely.
+An MCP server exposes the full skill library as callable tools and prompts. Any MCP-compatible client can discover all 165 skills, read their input requirements, call them with structured parameters, and receive typed outputs — removing the manual copy-paste step entirely.
 
 **Production URL:** `https://mcp-server-sigma-sooty.vercel.app/mcp`
 
@@ -85,7 +85,7 @@ The library is packaged as an Agent Skills 1.0–compliant plugin. The entry poi
 
 ### Progressive Disclosure
 
-The platform does not load all 107 skill files into context at once. Instead it uses metadata-first loading:
+The platform does not load all 165 skill files into context at once. Instead it uses metadata-first loading:
 
 1. **Discovery** — `registry.json` provides a lightweight index of all skills with descriptions, tags, and domain grouping. An orchestrator or the platform itself reads this to decide which skills are relevant.
 2. **Activation** — Only when a skill is selected (by the model, the user, or an orchestrator) is the full SKILL.md loaded into context.