social

TopNotchEuropeMeshAI

Cross-platform social workflow plugin and skills pack for Claude Code and Codex.

What This Repository Ships

• Claude plugin: social
• Claude marketplace metadata: .claude-plugin/marketplace.json
• Codex skills pack: skills/
• Canonical social workflow source: social/

The social workflow turns a repo into:

• a reusable project brief (.social/project-brief.json),
• Instagram content and publishing actions,
• LinkedIn draft/publish workflows,
• an 8-slide fundraising pitch deck draft.

Repository Layout

.
├── .claude-plugin/
│   ├── marketplace.json
│   └── plugin.json
├── .mcp.json
├── .env.example
├── codex/
│   └── INSTALL.md
├── scripts/
│   ├── run_with_env.sh
│   ├── validate_env.sh
│   ├── sync_social_skills.sh
│   ├── validate_universal_packages.sh
│   └── build_universal_packages.sh
├── skills/                  # Codex-distributed mirror (generated)
└── social/                  # Canonical source for commands, MCP, and skill content
    ├── .claude-plugin/plugin.json
    ├── .mcp.json
    ├── commands/
    ├── scripts/
    ├── server/
    └── skills/

Install

Claude Code

Add this repo as a marketplace source:

claude --plugin add-marketplace barteksad/top-notch-europe-mesh-ai

Install the plugin from that marketplace:

claude --plugin install social@topnotch-social-marketplace

For local development, install directly from a local clone:

claude --plugin install --local /absolute/path/to/top-notch-europe-mesh-ai

Codex

Tell Codex to fetch and follow:

https://raw.githubusercontent.com/barteksad/top-notch-europe-mesh-ai/main/codex/INSTALL.md

Or read local instructions in codex/INSTALL.md.

Local Prerequisites

• Python 3.10+
• jq, rsync, tar
• Python dependencies:

python3 -m pip install -r social/scripts/requirements.txt
python3 -m venv social/server/.venv
social/server/.venv/bin/pip install -r social/server/requirements.txt

Unified Environment Setup

This repo follows a single env model for both Claude and Codex:

1. Copy env template:

cp .env.example .env

2. Fill .env with real values.

3. Optionally add machine-specific overrides in .env.local.

4. Validate env values for your target workflow:

./scripts/validate_env.sh --load --profile all

Use narrower profiles when needed:

./scripts/validate_env.sh --load --profile instagram
./scripts/validate_env.sh --load --profile linkedin
./scripts/validate_env.sh --load --profile media

5. Start agents through the env launcher so they inherit values:

./scripts/run_with_env.sh claude
./scripts/run_with_env.sh codex

Important:

• .env and .env.local are gitignored.
• Env changes require restarting the agent process.
• MCP server env mapping is explicit in .mcp.json and social/.mcp.json.

Environment Variables

Variable	Required for
GEMINI_API_KEY	AI media generation
INSTAGRAM_ACCESS_TOKEN	Instagram publish
INSTAGRAM_USER_ID	Instagram publish
GITHUB_TOKEN	Upload local media before Instagram publish
GITHUB_REPO	Upload local media before Instagram publish
LINKEDIN_CLIENT_ID	LinkedIn CLI auth
LINKEDIN_CLIENT_SECRET	LinkedIn CLI auth/token exchange
LINKEDIN_CLI_REDIRECT_URI	LinkedIn CLI auth

Maintainer Workflow

Treat social/ as source of truth.

1. Edit plugin source files only under social/.
2. Mirror skills into skills/:

./scripts/sync_social_skills.sh --validate

3. Validate packaging and env template consistency:

./scripts/validate_universal_packages.sh

4. Build distributable bundles:

./scripts/build_universal_packages.sh

Generated artifacts are written to dist/.

Distribution and Release Checklist

1. Update version in:

• .claude-plugin/plugin.json
• social/.claude-plugin/plugin.json
• .claude-plugin/marketplace.json (metadata.version and plugin version)

2. Run:

./scripts/build_universal_packages.sh

3. Verify artifacts:

• dist/topnotch-social-claude-marketplace-v<version>.tar.gz
• dist/topnotch-social-codex-skills-v<version>.tar.gz
• compatibility aliases:
  • dist/topnotch-social-claude-marketplace.tar.gz
  • dist/topnotch-social-codex-skills.tar.gz
  • dist/topnotch-social-agent-skills.tar.gz

4. Tag and publish release on GitHub with built artifacts attached.

License

Apache-2.0. See LICENSE.