open-sourceror

name	open-sourceror
description	Prepare Claude Code skills, agents, or collections for open-source sharing on GitHub. Supports two modes: (1) Standalone repo creation, or (2) Marketplace integration into existing plugin repos. Use when: - "prepare for open source", "open source this skill" - "upload skill to github", "share this agent" - "add to marketplace", "add to robot-tools" - "create repo for skill", "package for sharing" - User has a skill directory, agent file, or collection to share
triggers	["prepare for open source","open source this skill","upload skill to github","share this agent","add to marketplace","create repo for skill","package for sharing"]

Open Sourceror

Prepare AI-related components (skills, agents, collections) for open-source sharing.

Mode Selection

Ask user which mode to use:

Mode	Use Case
Standalone	Create new private repo for the component
Marketplace	Add to existing multi-plugin repo (e.g., robot-tools)

Default to Marketplace if user mentions an existing repo name or "add to".

---

Mode 1: Standalone Repository

Create a new private repository for the component.

1. Gather Information

Ask user for:

• Source path: Local path to skill directory, agent file, or collection
• Repo name: GitHub repository name (default: component name)
• Description: One-line repo description
• Author name: For LICENSE copyright (default: detect from git config user.name)

Detect automatically:

• GitHub username: gh api user --jq '.login'
• Current year: For LICENSE copyright
• Component type: Skill (has SKILL.md), Agent (.md with agent frontmatter), or Mixed

2. Detect Component Type

If source has SKILL.md → Single Skill
If source is .md file with "tools:" in frontmatter → Single Agent
If source has skills/ or agents/ subdirs → Mixed Collection
If source has multiple .md files with frontmatter → Mixed Collection

3. Create Repository

gh repo create {{REPO_NAME}} --private --description "{{DESCRIPTION}}"

4. Upload Files

cd /tmp && rm -rf {{REPO_NAME}}-upload
mkdir {{REPO_NAME}}-upload && cd {{REPO_NAME}}-upload
cp -r {{SOURCE_PATH}}/* .
git init -b main
git remote add origin https://github.com/{{OWNER}}/{{REPO_NAME}}.git
git add -A
git commit -m "feat: Initial upload of {{COMPONENT_NAME}}"
git push -u origin main

5. Generate README

Select template from references/readme-templates.md based on component type.

6. Add LICENSE

Copy assets/LICENSE template, replacing:

• {{YEAR}} → Current year
• {{AUTHOR}} → User-provided or detected author name

7. Push Documentation

cd /tmp && rm -rf {{REPO_NAME}}-docs
git clone https://github.com/{{OWNER}}/{{REPO_NAME}}.git {{REPO_NAME}}-docs
cd {{REPO_NAME}}-docs
# Write README.md and LICENSE
git add -A
git commit -m "docs: Add README and LICENSE"
git push

8. Report

✅ Repository created: https://github.com/{{OWNER}}/{{REPO_NAME}}

| Property | Value |
|----------|-------|
| Visibility | Private |
| Files | {{FILE_COUNT}} |
| README | ✓ |
| LICENSE | MIT with Commercial Restriction |

**Next step**: Make public when ready:
gh repo edit {{OWNER}}/{{REPO_NAME}} --visibility public

---

Mode 2: Marketplace Integration

Add component to an existing multi-plugin repository.

1. Gather Information

Ask user for:

• Source path: Local path to skill directory or agent file
• Plugin repo: GitHub repo URL or name (e.g., swannysec/robot-tools)
• Target toolkit: Which toolkit to add to (e.g., research-toolkit, workflow-toolkit)

If user has a default configured in ~/.config/open-sourceror/config.json, use that:

{
  "default_plugin_repo": "swannysec/robot-tools",
  "default_toolkit": null
}

Detect automatically:

• Component type: Skill or Agent
• Component name: From directory name or frontmatter

2. Validate Plugin Repo Structure

# Clone to temp
cd /tmp && rm -rf {{REPO_NAME}}-marketplace
gh repo clone {{PLUGIN_REPO}} {{REPO_NAME}}-marketplace
cd {{REPO_NAME}}-marketplace

Verify structure:

# Must have marketplace.json
test -f .claude-plugin/marketplace.json

# Must have target toolkit
test -d {{TARGET_TOOLKIT}}
test -f {{TARGET_TOOLKIT}}/.claude-plugin/plugin.json

3. Determine Target Directory

If Skill (has SKILL.md) → {{TARGET_TOOLKIT}}/skills/{{COMPONENT_NAME}}/
If Agent (.md file) → {{TARGET_TOOLKIT}}/agents/{{COMPONENT_NAME}}.md

4. Copy Component

# For skill
cp -r {{SOURCE_PATH}} {{TARGET_TOOLKIT}}/skills/

# For agent
cp {{SOURCE_PATH}} {{TARGET_TOOLKIT}}/agents/

5. Update Toolkit README

Read {{TARGET_TOOLKIT}}/README.md and add entry to appropriate table:

For Skills - add to Skills table:

| `{{SKILL_NAME}}` | {{DESCRIPTION}} |

For Agents - add to Agents table:

| `{{AGENT_NAME}}` | {{DESCRIPTION}} |

Also add:

• Trigger phrases to Usage section
• Example command to Examples section
• Any new requirements to Requirements section

6. Create Branch and Commit

git checkout -b feat/{{COMPONENT_NAME}}
git add {{TARGET_TOOLKIT}}/
git commit -m "feat({{TARGET_TOOLKIT}}): add {{COMPONENT_NAME}} {{TYPE}}"
git push -u origin feat/{{COMPONENT_NAME}}

7. Create Pull Request

gh pr create \
  --title "feat({{TARGET_TOOLKIT}}): add {{COMPONENT_NAME}} {{TYPE}}" \
  --body "## Summary

Add \`{{COMPONENT_NAME}}\` {{TYPE}} to {{TARGET_TOOLKIT}}.

{{DESCRIPTION}}

## Files Added

\`\`\`
{{FILE_LIST}}
\`\`\`

## Test plan

- [ ] Verify {{TYPE}} triggers on relevant phrases
- [ ] Test {{TYPE}} functionality"

8. Clean Up and Report

rm -rf /tmp/{{REPO_NAME}}-marketplace

Report:

✅ Pull request created: {{PR_URL}}

| Property | Value |
|----------|-------|
| Repository | {{PLUGIN_REPO}} |
| Toolkit | {{TARGET_TOOLKIT}} |
| Component | {{COMPONENT_NAME}} |
| Type | {{TYPE}} |
| Branch | feat/{{COMPONENT_NAME}} |

**Next step**: Review and merge the PR.

---

Toolkit Selection Guide

If user doesn't specify toolkit, suggest based on component purpose:

Component Purpose	Suggested Toolkit
Research, discovery, information gathering	research-toolkit
Security analysis, vulnerability scanning	security-toolkit
Code analysis, metrics, visualization	code-analysis-toolkit
Workflow automation, productivity, sharing	workflow-toolkit

---

Notes

• Always create branches for marketplace mode—never push directly to main
• Use conventional commit format: feat(toolkit): add component-name type
• Standalone repos are created as private—user controls visibility
• Verify target toolkit exists before proceeding
• Extract description from frontmatter for README updates