agent-sdk-go: a Go framework for multi-model agents with sessions, hooks, and evals

Ingenimax/agent-sdk-go is a Go framework for building agents that need to run beyond a single prompt-response loop. The README describes integrations with OpenAI, Anthropic, and Google Vertex AI (Gemini), a modular tool system, and memory options covering buffer, vector-based retrieval, and summarization. Full documentation is hosted separately at docs.goagents.dev, which matters for adoption: the repository README is a feature map, not a substitute for the guides and reference. The intended audience is Go teams already operating services who want agent orchestration inside their existing stack, rather than assembling memory, tool execution, and provider switching themselves. That positioning shows in the developer-facing pieces: declarative YAML agent and task definitions, a structured task framework for planning, approval, and multi-step execution, and a CLI that can be installed from pre-built binaries, via go install, or built from source. As a library, installation is a standard go get github.com/Ingenimax/agent-sdk-go. The stated prerequisite is Go 1.26+, with Redis optional