[skills] Add standalone community skills category

Author: justfinethanku

.github/ISSUE_TEMPLATE/bug-report.yml

@@ -11,7 +11,7 @@ body:
     id: contribution
     attributes:
       label: Which contribution?
-      description: The name and category (e.g., "recipes/email-history-import")
+      description: The name and category (e.g., "recipes/email-history-import" or "skills/panning-for-gold")
       placeholder: recipes/email-history-import
     validations:
       required: true

.github/ISSUE_TEMPLATE/feature-request.yml

@@ -12,7 +12,7 @@ body:
     attributes:
       label: What area does this relate to?
       options:
-        - An existing contribution (recipe, schema, dashboard, integration)
+        - An existing contribution (recipe, schema, dashboard, integration, skill)
         - Repo structure or templates
         - CI/CD or automation
         - Documentation

.github/ISSUE_TEMPLATE/recipe-submission.yml

@@ -1,5 +1,5 @@
 name: Recipe / Contribution Idea
-description: Propose a new recipe, schema, dashboard, or integration
+description: Propose a new recipe, schema, dashboard, integration, or skill
 labels: ["contribution-idea"]
 body:
   - type: markdown
@@ -17,6 +17,7 @@ body:
         - Schema (database extension)
         - Dashboard (frontend template)
         - Integration (MCP extension, webhook, capture source)
+        - Skill (reusable AI client skill or prompt pack)
         - Primitive (reusable concept guide)
         - Extension (curated learning path build)
         - Not sure

.github/PULL_REQUEST_TEMPLATE.md

@@ -5,6 +5,7 @@
 - [ ] Schema (`/schemas`)
 - [ ] Dashboard (`/dashboards`)
 - [ ] Integration (`/integrations`)
+- [ ] Skill (`/skills`)
 - [ ] Repo improvement (docs, CI, templates)
 
 ## What does this do?

.github/metadata.schema.json

@@ -2,7 +2,7 @@
   "$schema": "https://json-schema.org/draft/2020-12/schema",
   "$id": "https://github.com/NateBJones-Projects/OB1/blob/main/.github/metadata.schema.json",
   "title": "Open Brain Source Contribution Metadata",
-  "description": "Schema for metadata.json files in OB1 contributions (recipes, schemas, dashboards, integrations, primitives, extensions).",
+  "description": "Schema for metadata.json files in OB1 contributions (recipes, schemas, dashboards, integrations, skills, primitives, extensions).",
   "type": "object",
   "required": [
     "name",
@@ -28,7 +28,7 @@
     },
     "category": {
       "type": "string",
-      "enum": ["recipes", "schemas", "dashboards", "integrations", "primitives", "extensions"],
+      "enum": ["recipes", "schemas", "dashboards", "integrations", "skills", "primitives", "extensions"],
       "description": "Which contribution directory this belongs to."
     },
     "author": {

.github/workflows/auto-label.yml

@@ -35,6 +35,7 @@ jobs:
               'schemas/': 'schema',
               'dashboards/': 'dashboard',
               'integrations/': 'integration',
+              'skills/': 'skill',
               'primitives/': 'primitive',
               'extensions/': 'extension',
             };

.github/workflows/claude-issue-triage.yml

@@ -32,6 +32,7 @@ jobs:
                - Primitive proposal (new reusable concept guide)
                - Dashboard idea (new frontend template)
                - Integration idea (new MCP extension, webhook, or capture source)
+               - Skill idea (new reusable AI client skill or prompt pack)
                - Documentation improvement
                - Question (not a contribution, needs help)
 
@@ -46,13 +47,13 @@ jobs:
 
             4. **Suggest labels** — Based on classification, suggest which labels
                should be applied (recipe, schema, extension, primitive, dashboard,
-               integration, documentation, bug, question, good-first-issue).
+               integration, skill, documentation, bug, question, good-first-issue).
 
             5. **Helpful response** — Post a welcoming, constructive comment:
                - Acknowledge the idea
                - If they used the blank template, point them to the right
                  specific template for better structure
-               - If it's a recipe/schema/integration idea, link to
+               - If it's a recipe/schema/integration/skill idea, link to
                  CONTRIBUTING.md's "What Goes Where" section
                - If it looks like a duplicate, link the existing issue
                - If it's a question, try to answer or point to docs/FAQ

.github/workflows/ob1-review.yml

@@ -36,7 +36,7 @@ jobs:
 
           # Identify contribution folders (e.g., recipes/email-import/)
           # Exclude _template folders and top-level files
-          contrib_dirs=$(echo "$changed" | grep -E '^(recipes|schemas|dashboards|integrations|primitives|extensions)/' | grep -v '/_template/' | cut -d'/' -f1,2 | sort -u || true)
+          contrib_dirs=$(echo "$changed" | grep -E '^(recipes|schemas|dashboards|integrations|skills|primitives|extensions)/' | grep -v '/_template/' | cut -d'/' -f1,2 | sort -u || true)
           echo "contrib_dirs<<EOF" >> $GITHUB_OUTPUT
           echo "$contrib_dirs" >> $GITHUB_OUTPUT
           echo "EOF" >> $GITHUB_OUTPUT
@@ -83,7 +83,7 @@ jobs:
           # ─── Rule 1: Folder structure ───
           rule1_pass=true
           bad_files=""
-          allowed_pattern='^(recipes|schemas|dashboards|integrations|primitives|extensions|docs|resources|\.github|\.gitkeep|\.gitignore|\.markdownlint)'
+          allowed_pattern='^(recipes|schemas|dashboards|integrations|skills|primitives|extensions|docs|resources|\.github|\.gitkeep|\.gitignore|\.markdownlint)'
           while IFS= read -r f; do
             [ -z "$f" ] && continue
             if ! echo "$f" | grep -qE "$allowed_pattern"; then
@@ -286,6 +286,13 @@ jobs:
                   rule6_pass=false
                 fi
                 ;;
+              skills)
+                has_skill=$(echo "$dir_files" | grep -iE '(^|/)(SKILL\.md|[^/]+[.-]skill\.md)$' || true)
+                if [ -z "$has_skill" ]; then
+                  rule6_detail="${rule6_detail}  - \`$dir\`: skills must contain a plain-text skill file (\`SKILL.md\`, \`*.skill.md\`, or \`*-skill.md\`)\n"
+                  rule6_pass=false
+                fi
+                ;;
               primitives)
                 # Primitives must have substantial README (200+ words)
                 if [ -f "$dir/README.md" ]; then
@@ -319,10 +326,10 @@ jobs:
           fi
 
           # ─── Rule 7: PR format ───
-          if echo "$PR_TITLE" | grep -qE '^\[(recipes|schemas|dashboards|integrations|primitives|extensions|docs)\] '; then
+          if echo "$PR_TITLE" | grep -qE '^\[(recipes|schemas|dashboards|integrations|skills|primitives|extensions|docs)\] '; then
             pass_check "PR format" "Title follows \`[category] Description\` format"
           else
-            fail_check "PR format" "PR title must start with \`[recipes]\`, \`[schemas]\`, \`[dashboards]\`, \`[integrations]\`, \`[primitives]\`, \`[extensions]\`, or \`[docs]\` followed by a space and description"
+            fail_check "PR format" "PR title must start with \`[recipes]\`, \`[schemas]\`, \`[dashboards]\`, \`[integrations]\`, \`[skills]\`, \`[primitives]\`, \`[extensions]\`, or \`[docs]\` followed by a space and description"
           fi
 
           # ─── Rule 8: No binary blobs ───

CLAUDE.md

@@ -4,7 +4,7 @@ This file helps AI coding tools (Claude Code, Codex, Cursor, etc.) work effectiv
 
 ## What This Repo Is
 
-Open Brain is a persistent AI memory system — one database (Supabase + pgvector), one MCP protocol, any AI client. This repo contains the extensions, recipes, schemas, dashboards, and integrations that the community builds on top of the core Open Brain setup.
+Open Brain is a persistent AI memory system — one database (Supabase + pgvector), one MCP protocol, any AI client. This repo contains the extensions, recipes, schemas, dashboards, integrations, and skills that the community builds on top of the core Open Brain setup.
 
 **License:** FSL-1.1-MIT. No commercial derivative works. Keep this in mind when generating code or suggesting dependencies.
 
@@ -17,8 +17,9 @@ recipes/        — Standalone capability builds. Open for community contributio
 schemas/        — Database table extensions. Open.
 dashboards/     — Frontend templates (Vercel/Netlify). Open.
 integrations/   — MCP extensions, webhooks, capture sources. Open.
+skills/         — Reusable AI client skills and prompt packs. Open.
 docs/           — Setup guides, FAQ, companion prompts.
-resources/      — Claude Skill, companion files.
+resources/      — Official companion files and packaged exports.
 ```
 
 Every contribution lives in its own subfolder under the right category and must include `README.md` + `metadata.json`.
@@ -33,7 +34,7 @@ Every contribution lives in its own subfolder under the right category and must
 
 ## PR Standards
 
-- **Title format:** `[category] Short description` (e.g., `[recipes] Email history import via Gmail API`)
+- **Title format:** `[category] Short description` (e.g., `[recipes] Email history import via Gmail API`, `[skills] Panning for Gold standalone skill pack`)
 - **Branch convention:** `contrib/<github-username>/<short-description>`
 - **Commit prefixes:** `[category]` matching the contribution type
 - Every PR must pass the automated review checks in `.github/workflows/ob1-review.yml` before human review

CONTRIBUTING.md

@@ -30,18 +30,20 @@ This is a first-class path, not a workaround. Some of the best contributions com
 | `schemas/` | Database table extensions and metadata schemas | CRM contacts table, taste tracker, reading list schema | Open |
 | `dashboards/` | Frontend templates for Vercel/Netlify hosting | Knowledge dashboard, weekly review, mobile capture UI | Open |
 | `integrations/` | MCP extensions, webhooks, capture sources | Discord bot, email handler, browser extension, calendar sync | Open |
+| `skills/` | Reusable AI client skills and prompt packs | Meeting triage assistant, code review protocol, transcript processor | Open |
 
-### Extensions vs Primitives vs Recipes
+### Extensions vs Primitives vs Recipes vs Skills
 
 - **Extensions** are curated, ordered builds that form a progressive learning path. Each teaches new concepts through practical use. They include database schemas, MCP server code, and step-by-step instructions. If you want to propose a new extension, [open an issue](../../issues/new?template=extension-submission.yml) first.
 - **Primitives** are reusable concept guides that get referenced by multiple extensions. They teach a pattern (like RLS or shared access) once, so extensions can link to them instead of re-explaining. A primitive should be referenced by at least 2 extensions. [Propose one here](../../issues/new?template=primitive-submission.yml).
 - **Recipes** are standalone builds — they add a capability without being part of the learning path. No ordering, no prerequisites beyond a working Open Brain. Open for community contributions.
+- **Skills** are standalone agent behaviors packaged as plain-text prompt/skill files. They are smaller than recipes: no full build required, just a reusable behavior you can install into Claude Code, Codex, Cursor, or a similar client. Open for community contributions.
 
 Not sure where yours fits? Open a discussion issue first.
 
 ## Required Files
 
-Every contribution lives in its own subfolder under the right category (e.g., `recipes/my-cool-recipe/`) and must include:
+Every contribution lives in its own subfolder under the right category (e.g., `recipes/my-cool-recipe/` or `skills/my-cool-skill/`) and must include:
 
 - **`README.md`** — What it does, prerequisites, step-by-step setup, expected outcome, troubleshooting
 - **`metadata.json`** — Structured metadata (see template below)
@@ -130,6 +132,13 @@ grant select, insert, update, delete on table public.your_table to service_role;
 **Primitives** additionally require:
 - **"Extensions That Use This"** section listing which extensions reference this primitive
 
+**Skills** additionally require:
+- At least one plain-text skill artifact (`SKILL.md`, `*.skill.md`, or `*-skill.md`)
+- **"Supported Clients"** section listing which AI tools the skill is designed for
+- **"Installation"** steps showing exactly where the file goes
+- **"Trigger Conditions"** explaining when the skill should fire
+- Plain-text, reviewable source files only. Do not submit zipped/exported skill bundles in `skills/`.
+
 Check the `_template/` folder in each category for a starter README. The extension template contains HTML comments with detailed instructions for both human contributors and AI assistants.
 
 ## metadata.json
@@ -201,6 +210,7 @@ Example for an extension:
 - Example: `[schemas] CRM contacts table with interaction tracking`
 - Example: `[extensions] Household Knowledge Base`
 - Example: `[primitives] Row Level Security guide`
+- Example: `[skills] Panning for Gold standalone skill pack`
 
 **Description must include:**
 - What the contribution does
@@ -239,13 +249,13 @@ Non-code contributions count at every level. Testing recipes, mentoring non-tech
 
 Every PR is checked against these rules. All must pass before human review.
 
-1. **Folder structure** — Contribution is in the correct category directory (`recipes/`, `schemas/`, `dashboards/`, `integrations/`, `primitives/`, `extensions/`)
+1. **Folder structure** — Contribution is in the correct category directory (`recipes/`, `schemas/`, `dashboards/`, `integrations/`, `skills/`, `primitives/`, `extensions/`)
 2. **Required files** — Both `README.md` and `metadata.json` exist in the contribution folder
 3. **Metadata valid** — `metadata.json` parses as valid JSON and contains all required fields
 4. **No credentials** — No API keys, tokens, passwords, or secrets in any file
 5. **SQL safety** — No `DROP TABLE`, `DROP DATABASE`, `TRUNCATE`, or unqualified `DELETE FROM`. No modifications to core `thoughts` table columns (adding columns is fine, altering/dropping existing ones is not)
-6. **Category-specific artifacts** — `recipes/` have code or detailed instructions, `schemas/` have SQL files, `dashboards/` have frontend code or `package.json`, `integrations/` have code files, `primitives/` have substantial READMEs (200+ words), `extensions/` have both SQL and code files
-7. **PR format** — Title starts with `[recipes]`, `[schemas]`, `[dashboards]`, `[integrations]`, `[primitives]`, or `[extensions]`
+6. **Category-specific artifacts** — `recipes/` have code or detailed instructions, `schemas/` have SQL files, `dashboards/` have frontend code or `package.json`, `integrations/` have code files, `skills/` have at least one plain-text skill file, `primitives/` have substantial READMEs (200+ words), `extensions/` have both SQL and code files
+7. **PR format** — Title starts with `[recipes]`, `[schemas]`, `[dashboards]`, `[integrations]`, `[skills]`, `[primitives]`, or `[extensions]`
 8. **No binary blobs** — No files over 1MB, no `.exe`, `.dmg`, `.zip`, `.tar.gz`
 9. **README completeness** — Contribution README includes Prerequisites, step-by-step instructions, and expected outcome sections
 10. **Primitive dependencies** — If a contribution declares `requires_primitives`, the primitives must exist in the repo and be linked in the README