FlashAgent

FlashAgent is a coding agent distributed as a single native binary for the terminal, designed primarily for local models. According to the project page, it "finds the model your local server has loaded, and tells you whether that model can actually drive tools." It is described as being in beta (b477) and actively developed by a solo maintainer, who also uses the tool on its own codebase.

The agent connects to local model servers such as LM Studio, Ollama, llama.cpp, and vLLM, as well as cloud APIs including Anthropic, OpenAI, Gemini, and OpenRouter. The page states that it "reads the loaded model, its context window and its reasoning presets from LM Studio, Ollama, llama.cpp or vLLM," and that a `/provider` command can move a conversation between a local server and a cloud one without restarting. Installation is available via shell scripts for Linux, macOS, and Windows, as well as packages for several platforms.

A central feature is tool-calling support and testing. FlashAgent provides "Native tool calls, a recovery parser for calls written as text, JSON repair, and a test that shows which models keep up." The `--tool-test` command runs eight scenarios to evaluate whether a model can actually invoke tools, and the page notes that "Small models do better since b450, which tells them how to call tools, repairs their commonest argument mistakes and reminds them once of a step left undone." The binary is described as approximately 16 MB with about 10 MB of memory when idle, and it includes four permission modes and approval cards that show the exact command or diff before execution.