datarobot-agent-assist

What problem does it solve?

Teams need a repeatable way to design DataRobot agent specifications, validate them before writing code, and then move from an implemented spec to a deployed DataRobot custom application with fewer configuration mistakes.

Core Features & Use Cases

• Agent design assistance with spec generation: Clarifies requirements (including prompts, tools, and optional frontend needs) and iteratively writes a valid agent_spec.md in YAML.
• Optional dress-rehearsal simulation before coding: Runs an end-to-end rehearsal loop via the DataRobot LLM Gateway to test tool calls and system prompt behavior before implementation.
• Workflow-guided coding + deployment readiness: Drives users through prerequisites (Git, Python 3.11+, DataRobot CLI), template preparation, framework selection, dependency checks, and then points to AGENTS.md for deployment.

Use case example: You want to build a DataRobot agent for “agent spec first, then implement,” including a “multi-page/custom frontend” requirement, and you want to simulate tool interactions to catch spec issues early.

Quick Start

Use the datarobot-agent-assist skill to design and validate your DataRobot agent by starting from your agent_spec idea and selecting Design an AI agent in the skill menu.