Looking through the thread for open source projects explicitly mentioned as AI-assisted, here's what the source actually documents: **Clearly AI-assisted and open source:** - **aep-conformance-test** (thegagne): Used Claude to write conformance tests for an API spec. Source says he "inventoried every MUST, SHOULD, and MAY" in the spec, then had Claude generate tests. GitHub link provided. - **reviewflow** (Luyanda): VSCode/Cursor extension for code reviews, explicitly open source on GitHub, though the source doesn't detail what AI assistance went into building it. - **trystero** (dmotz): P2P apps framework—the source describes it but doesn't specify AI involvement in development. - **goldseam** (dataviz1000): Self-healing test selectors for Cypress using Claude Code or self-hosted models. Open source on GitHub. - **posipaki** (Muromec): Actor library in TypeScript; mentions using DeepSeek for development but unclear how much of the actual codebase is AI-generated. **Mentioned as using LLMs but less clear on open source status:** - **ffsubsync, ipyflow, pipescript, pycograd** (smacke): Multiple projects now browser-runnable via pyodide, developed with LLM assistance, but the source doesn't confirm they're open source (though links suggest they might be). The source doesn't say how many projects total had AI assistance during development—that detail isn't available. Many mention using Claude Code or similar tools, but don't commit to releasing code publicly.