docs(skill): README + CONTRIBUTING + PR template for skill submissions

Author: KylinMountain

.github/PULL_REQUEST_TEMPLATE/skill_submission.md

@@ -0,0 +1,22 @@
+## Skill submission
+
+- **Name:** `<skill-name>`
+- **What it does (one sentence):**
+- **Source material:**
+
+### Checklist
+
+- [ ] I have the right to redistribute the source material this skill is
+      derived from (public, my own work, properly licensed, or fair-use
+      methodology only — not bulk-copied copyrighted text).
+- [ ] `SKILL.md` has valid YAML frontmatter with `name:` and `description:`
+      (description ≤ 1024 chars).
+- [ ] All `[[references/...]]` links resolve.
+- [ ] I added an entry under `plugins[0].skills` in
+      `.claude-plugin/marketplace.json`.
+- [ ] I installed and tested this skill locally in at least one agent CLI
+      (Claude Code / Codex CLI / Gemini CLI / Cursor).
+
+### Notes for reviewers
+
+<anything else maintainers should know>

CONTRIBUTING.md

@@ -0,0 +1,71 @@
+# Contributing to OpenKB
+
+Thanks for your interest in OpenKB! There are two ways to contribute: code (bug
+fixes, new features) and skills (compiled knowledge artifacts).
+
+## Code contributions
+
+Standard GitHub workflow:
+
+1. Fork the repo and create a feature branch off `main`.
+2. Add tests for any new behaviour. Run `uv run pytest` to verify they pass.
+3. Open a PR with a clear description and link to any related issues.
+4. Address review feedback. A maintainer will merge once everything looks good.
+
+## Submitting a skill
+
+OpenKB doubles as a registry for compiled-knowledge skills. The `skills/`
+directory in this repo hosts the official `openkb` skill plus
+community-contributed skills installable via:
+
+```bash
+npx skills@latest add VectifyAI/OpenKB
+```
+
+To submit your skill:
+
+### 1. Compile and test locally
+
+```bash
+cd <your-kb>
+openkb skill new <skill-name> "<one-sentence intent>"
+cp -r output/skills/<skill-name> ~/.claude/skills/
+# Verify the skill activates in a Claude Code session on a relevant question
+```
+
+### 2. Open a PR against this repo
+
+1. Fork `VectifyAI/OpenKB`.
+2. Copy your skill into the fork at `skills/<your-skill-name>/`.
+3. Add a new entry under `plugins[0].skills` in `.claude-plugin/marketplace.json`:
+   ```json
+   "./skills/<your-skill-name>"
+   ```
+4. Open a PR with the "Skill submission" template (auto-applied when you create
+   the PR).
+
+### 3. Review
+
+A maintainer will review your submission against this checklist:
+
+- `SKILL.md` has a valid YAML frontmatter with `name:` and `description:`
+  (description ≤ 1024 characters).
+- The `description` is specific (not "a skill about X").
+- All `[[references/...]]` links resolve to files you actually included.
+- Skill name doesn't clash with an existing entry.
+- No obvious copyright-infringing content (large verbatim copies of recent
+  copyrighted books or papers without authorisation).
+
+You're responsible for confirming you have rights to redistribute the source
+material your skill is derived from. The PR template includes a checkbox for
+this.
+
+### What makes a good community skill
+
+- **Methodology-focused**, not bulk-copied content. "How to reason about X"
+  beats "the full text of X".
+- **Specific scope** so the `description:` field can be precise enough that
+  loading agents pick it up at the right time.
+- **Tested in at least one agent CLI** before submission.
+
+Thanks for sharing what you've compiled!

README.md

@@ -150,6 +150,7 @@ A single source might touch 10-15 wiki pages. Knowledge accumulates: each docume
 |---|---|
 | `openkb init` | Initialize a new knowledge base (interactive) |
 | <code>openkb&nbsp;add&nbsp;&lt;file_or_dir_or_URL&gt;</code> | Add documents and compile to wiki. URL ingest auto-detects PDF (saved as `.pdf` → PageIndex / markitdown) vs HTML (trafilatura main-content extract → `.md`) |
+| <code>openkb&nbsp;skill&nbsp;new&nbsp;&lt;name&gt;&nbsp;"&lt;intent&gt;"</code> | Compile a skill from this KB's wiki into `output/skills/<name>/` and update the marketplace manifest |
 | <code>openkb&nbsp;remove&nbsp;&lt;doc&gt;</code> | Remove a document and clean up its wiki pages, images, registry, and PageIndex state (use `--dry-run` to preview, `--keep-raw` / `--keep-empty-concepts` to retain artifacts) |
 | <code>openkb&nbsp;query&nbsp;"question"</code> | Ask a question over the knowledge base (use `--save` to save the answer to `wiki/explorations/`) |
 | `openkb chat` | Start an interactive multi-turn chat (use `--resume`, `--list`, `--delete` to manage sessions) |
@@ -161,6 +162,46 @@ A single source might touch 10-15 wiki pages. Knowledge accumulates: each docume
 
 <!-- | `openkb lint --fix` | Auto-fix what it can | -->
 
+### Skills — compile your wiki into a redistributable skill
+
+Once you have a populated wiki, you can compile a subset of it into an
+**Anthropic Skill** — a portable folder that Claude Code, Codex CLI,
+Gemini CLI, and Cursor all know how to load.
+
+```bash
+openkb skill new karpathy-thinking \
+  "Reason about transformers and attention in Karpathy's style"
+```
+
+This produces `output/skills/karpathy-thinking/` with `SKILL.md`,
+optional `references/`, and an auto-updated
+`.claude-plugin/marketplace.json` for distribution.
+
+**Install locally:**
+
+```bash
+cp -r output/skills/karpathy-thinking ~/.claude/skills/
+```
+
+**Share with others:**
+
+Push your KB directory to GitHub, then anyone can install all your skills with one command:
+
+```bash
+npx skills@latest add <your-org>/<your-repo>
+```
+
+You can also iterate inside chat:
+
+```
+/skill new karpathy-thinking "Reason about transformers like Karpathy"
+[generation streams]
+> description is too generic, make it about transformer implementations specifically
+[agent edits SKILL.md frontmatter in place]
+```
+
+See [CONTRIBUTING.md](CONTRIBUTING.md) for how to submit your compiled skill back to the OpenKB community registry.
+
 ### Interactive Chat
 
 `openkb chat` opens an interactive chat session over your wiki knowledge base. Unlike the one-shot `openkb query`, each turn carries the conversation history, so you can dig into a topic without re-typing context.