Tracked 179 · Curated 7
The popular Neovim plugin Nvim-treesitter, which has accumulated over 13,000 stars on GitHub, has been officially archived. Further details regarding the project status can be found on the project's GitHub discussion page.
OpenAI has updated the pricing structure for Codex, shifting from a per-message billing model to one based on API token usage. This transition aligns the cost calculation directly with the volume of tokens processed.
scan-for-secrets 0.2 has been released, introducing streaming results for the CLI tool to improve efficiency when scanning large directories. Key updates include support for multiple directory scanning, a new -f/--file option for scanning specific files, a -v/--verbose flag, and several new functions in the Python API, such as scan_directory_iter, scan_file, and scan_file_iter.
To maintain workflow efficiency, Gemini CLI users are advised to keep hooks fast. Because hooks run synchronously, any script delays will directly impact the agent's response time. Google suggests using parallel operations and caching for resource-intensive tasks to ensure optimal performance.
The Syntaqlite Playground is now live, providing a web-based interface for the syntaqlite library. By leveraging WebAssembly and Pyodide, the tool allows users to test various features, including SQL formatting, AST parsing, validation, and tokenization, directly within their browsers. The playground offers a user-friendly UI for exploring query structures and diagnostic feedback.
Claude Island is an open-source tool designed for the macOS Dynamic Island that acts as a notification manager for Claude Code. It addresses the friction of constantly switching back to the terminal to approve file operations or command executions by moving these prompts directly to the MacBook's notch area. This integration streamlines the workflow for developers using Claude Code, and a well-maintained fork is available for users.
Google Maps has integrated Gemini AI, allowing users to leverage the chatbot for itinerary planning. A hands-on test reveals that the feature performs surprisingly well, successfully suggesting personalized locations based on specific user requests and uncovering hidden gems that were previously unknown to the user.