GitHub Repo Briefing
free-claude-code
A launcher for terminal, desktop, IDE, and Discord workflows with Claude Code, Codex, Pi, OpenCode, and local model providers. It includes server and client commands for connecting those tools.
A first test
Install it in a test account and run fcc-server with one provider. Use fcc-claude on a read-only repository, then check the provider, token path, and command log before using a paid account.
ml-intern
An ML engineering agent that researches papers, writes code, and works with Hugging Face models, datasets, documentation, and compute. It can run interactively or from one headless prompt.
A first test
Clone it, run uv sync, and start with a local or sandboxed prompt such as inspect one dataset. Set only the tokens it needs and review every generated training or Hub action.
skills
A collection of small engineering skills for planning, implementation, testing, review, simplification, and release work. The files are editable and can be installed as a managed Claude Code plugin.
A first test
Install one skill set, not both installation methods. Run its planning and testing commands on a small change, then keep only the files that improve your actual review evidence.
hackingtool
A menu of more than 200 security tools across recon, OSINT, web, wireless, forensics, and post-exploitation categories, with an AI layer that suggests commands.
A first test
Use it only on systems that you own or are authorised to test. Pick one local lab target, run one tool manually first, and save the command and permission record with the result.
andrej-karpathy-skills
multica-ai/andrej-karpathy-skills
A single CLAUDE.md file that turns four coding rules into agent instructions: think before coding, keep the design simple, make small changes, and work toward a clear goal.
A first test
Copy the file into a disposable project and give the agent one small bug. Check whether it asks for the goal first and changes only the files that the fix needs.
claude-context
An MCP plugin that uses semantic search to find relevant code across a large repository and provide that context to Claude Code or another coding agent.
A first test
Install the Node 20+ package in a disposable clone and index it. Ask for one symbol's callers, compare the result with rg, and allow write tools only after the search is accurate.
FinceptTerminal
Fincept-Corporation/FinceptTerminal
An AGPL desktop finance application with market analytics, investment research, and economic data tools. The public repository is the free edition for learning and academic use.
A first test
Build the free edition and load a small set of historical data. Check the data source and date for every chart; do not use the output as a trading instruction without independent review.
ds2api
A Go middleware service that converts a DeepSeek web conversation interface into OpenAI, Claude, and Gemini-compatible API formats, with a React administration interface.
A first test
Run it locally with the documented Docker or Go path and send one test request. Inspect the request translation and terms of service before connecting a real account or client traffic.
Pixelle-Video
A web-based short-video pipeline that turns a topic into a script, images or video, voice narration, music, subtitles, and a final composition.
A first test
Use a non-client topic and one short output. Check the generated script, media licenses, voice, and subtitle timing before you reuse any asset in a public campaign.
Open-Generative-AI
Anil-matcha/Open-Generative-AI
A self-hosted image and video generation studio that connects to many model providers through a web interface. It is released under the MIT license and supports configurable model access.
A first test
Run it with a local test configuration and generate one image from a prompt you wrote. Record the provider, model, input, and output license before using the result commercially.
cua
Infrastructure, SDKs, and benchmarks for computer-use agents that control full macOS, Linux, and Windows desktops in sandboxes.
A first test
Start one sandbox and run the smallest desktop benchmark. Measure startup, screen access, file access, and cleanup, then keep the agent on a test account until those boundaries are clear.
GenericAgent
A small self-evolving agent with browser, terminal, file, keyboard, screen, and Android tools. It includes both a terminal interface and a web interface for testing those actions.
A first test
Create a uv environment and run the TUI against a throwaway folder. Ask it to list and edit one test file, then inspect every command before you allow a wider tool scope.