connectsafely-cookbook GitHub Details, Stars and Alternatives | OpenRepoFinder
ConnectSafelyAI / repository
connectsafely-cookbook
Open-source LinkedIn API cookbook: Claude Skills, MCP servers, n8n & Make.com workflows, and Python/TypeScript snippets for AI agents, prospecting, and outreach automation.
A transparent discovery signal based on current public GitHub metadata.
Recent activity35% weight
72
Community adoption25% weight
0
Maintenance state20% weight
100
License clarity10% weight
100
Project information10% weight
100
This score does not audit code, security, maintainers, documentation quality, or suitability. Verify the repository and its current documentation before adoption.
README preview
ConnectSafely Cookbook — LinkedIn API Examples, Claude Skills & MCP Recipes
Open-source recipes for the ConnectSafely.ai LinkedIn API: ready-to-run Claude Skills, Model Context Protocol (MCP) examples, n8n workflows, and code snippets for LinkedIn automation, outreach, lead research, and AI agents.
A growing library of runnable LinkedIn automation examples built on the ConnectSafely LinkedIn API. Every recipe runs end-to-end with only a single API key — no databases, no extra services. Designed for:
AI agent builders who need a LinkedIn tool surface for Claude, GPT, or open-source agent frameworks.
Growth and RevOps engineers who automate prospecting, enrichment, and outreach.
Developers integrating LinkedIn data into CRMs, lead-scoring pipelines, or recruiting workflows.
Code is plain Python or TypeScript with stdlib + requests/fetch. Skip the SDK churn.
Quickstart
Run the reference skill against any public LinkedIn profile in under a minute:
git clone https://github.com/ConnectSafelyAI/connectsafely-cookbook.git
cd connectsafely-cookbook
cp .env.example .env # then paste your CONNECTSAFELY_API_KEY
cd skills/prospect-researcher
# macOS / most Linux (stock Python ships pip3, not pip):
python3 -m pip install -r requirements.txt
python3 example.py https://www.linkedin.com/in/williamhgates/
# If your system has 'pip' and 'python' on PATH (Windows, Conda, venvs, some Linux):
pip install -r requirements.txt
python example.py https://www.linkedin.com/in/williamhgates/
The script auto-loads .env from the skill folder or any parent — a single .env at the repo root works for every skill. Real shell environment variables (export CONNECTSAFELY_API_KEY=…) always take priority over .env if both are set.
Both pip variants do the same thing.python3 -m pip is the portable form that always works; pip / python are the shorthand that exist on systems where Python is configured for them. Pick whichever your terminal recognizes.
You'll get a Markdown outreach brief on stdout. See the full sample output.
Claude Skills
Reusable Claude Skills — drop-in agent capabilities you can use from Claude Desktop, Claude Code, or the Anthropic API.
Auto-comment growth loops — comment on creators in your niche from a managed LinkedIn account.
Personalized outreach at scale — generate openers grounded in the prospect's recent posts.
Warm-intro discovery — find first-degree connections at a target company.
Each is one or two API calls. The skills/ folder turns them into copy-paste agent capabilities.
FAQ
What is the ConnectSafely API?
A managed LinkedIn API that wraps the most common automation actions — fetching profiles, sending messages, posting, commenting, searching — behind a single bearer-token endpoint. Your LinkedIn account stays connected through ConnectSafely's safety layer (rate-limit pacing, warmup, error recovery).
Is there a free tier?
The product has a free plan, but it is scoped to post-boosting only and does not include API access. API access requires the Ultimate Outreach plan at $10/month. New accounts receive a trial. See docs/pricing.md.
Which LinkedIn endpoints are supported?
Profiles, posts (read, comment, react, repost, search), messaging (DM and InMail), connection requests, follow/unfollow, people and company search, groups, jobs, account analytics, and Sales Navigator threads. Full reference at connectsafely.ai/docs.
Can I use this with Claude, ChatGPT, or LangChain?
Yes. The skills here ship as Claude Skills, but the underlying scripts are plain Python — wire them into LangChain tools, OpenAI function-calling, or any agent framework. The mcp/ folder shows the MCP integration path for Claude Desktop and Cursor.
How do I avoid getting rate-limited?
Read docs/rate-limits.md. Every response includes X-RateLimit-* headers — pace off those rather than guessing. Hardest limits: 90 connection requests/week and 120 unique profile fetches/day per LinkedIn account.
What if a recipe stops working?
Open a broken-skill issue. CI runs every skill weekly so we catch most drift early.
Contributing
Built something useful? Open a PR. The quality bar:
Runs end-to-end with only an API key
Has a committed sample output
Links every endpoint it calls back to the API docs
