This daily update tracks GitHub repositories experiencing the highest relative momentum over a rolling seven-day window, ranking projects by percentage star growth rather than absolute volume to surface emerging tools before they reach mainstream saturation. This week’s surge is led by Google’s Elephant Goldfish Model, which saw a 2,072% increase in stars by introducing specialized subagents for software design, alongside OpenBot’s 426% growth in the local-first workspace sector. The data highlights a concentrated interest in specialized production pipelines, such as Blender Skills’ 301% gain in 3D modeling automation, and infrastructure-level security tools like ThinkWatch Lite, which grew 240% by offering local API key redaction and cost tracking.
| # | Repository | Stars Gained | Gain % | Overview |
|---|---|---|---|---|
| 1 | google/elephant-goldfish-model | 6342 | 2072.5% | An open-source plugin for AI coding agents that uses specialized subagents to brainstorm, draft, and stress-test technical specifications before implementation. |
| 2 | nightly-labs/openbot | 3442 | 426% | A local-first desktop workspace for managing persistent AI teammates across multiple providers like Claude, Gemini, and Grok. |
| 3 | arjun988/blender-skills | 1188 | 301.5% | A specialized library of 94 AI skills that enables LLM agents to execute complex 3D modeling, texturing, and animation workflows directly within Blender via MCP. |
| 4 | ThinkWatchProject/ThinkWatch-Lite | 296 | 240.7% | A local gateway for AI clients that provides centralized management, security filtering, and cost tracking for multiple LLM providers. |
| 5 | C4rohan/Lampline | 380 | 226.2% | An open-source framework for building, testing, and running self-hosted voice agents with automated AI-driven call evaluations. |
Elephant Goldfish Model (google/elephant-goldfish-model )
The Elephant Goldfish Model (EGM) is an open-source plugin designed to enhance the software design process through specialized AI roles. It utilizes ‘Elephant’ agents to maintain project context for drafting specifications and ‘Goldfish’ agents to perform zero-context reviews to expose unstated assumptions. Developers use EGM within environments like Claude Code or VS Code to brainstorm ideas, generate design documents, and run drift reports. This workflow ensures architectural trade-offs are resolved and specifications are rigorously vetted before any code is written.
OpenBot (nightly-labs/openbot )
OpenBot is a desktop workspace designed to host persistent AI agents with individual workspaces and conversation histories. It supports a wide range of providers, including Claude Code, Gemini, and Grok, while offering local-first data management and agent-to-agent messaging. Users can leverage features like per-agent context monitoring, embedded browsers, and computer use capabilities to automate complex workflows. It is primarily intended for developers and power users who require a centralized, privacy-conscious environment to orchestrate multiple AI agents.
Blender Skills (arjun988/blender-skills )
Blender Skills provides a comprehensive suite of 94 specialized instructions designed to transform AI agents like Claude Code and Cursor into technical art teams. By leveraging the Model Context Protocol (MCP), these skills allow users to automate entire production pipelines, from hard-surface modeling and retopology to complex lighting and scene assembly. The repository includes orchestrated workflows for specific genres, such as horror or sci-fi, and supports reference image matching to ensure visual accuracy. It is primarily intended for 3D artists and technical directors looking to integrate generative AI into professional Blender production environments.
ThinkWatch Lite (ThinkWatchProject/ThinkWatch-Lite )
ThinkWatch Lite acts as a local proxy for AI clients like Claude Code, Cursor, and Aider, allowing users to switch between different LLM upstreams without reconfiguring individual tools. It provides critical security features, including the ability to redact sensitive data like API keys and bank card numbers before they leave the machine, and blocking dangerous tool calls. The platform also offers detailed observability, enabling users to track request costs, compare model performance across providers, and audit full request histories. It is designed for developers and privacy-conscious users who need centralized control over their AI workflows and expenditures.
Lampline (C4rohan/Lampline )
Lampline is an open-source platform designed for developing and testing conversational voice agents locally or on-premises. It enables developers to build agents using YAML and Markdown, integrate with over 20 model providers, and run automated evaluations where an AI caller tests real phone lines. The tool supports importing and exporting Google Dialogflow CX agents and integrates with telephony providers like Twilio and Telnyx. It is primarily used by engineers to ensure voice agent reliability through rigorous, automated testing and cost-effective, self-hosted deployment.