What problem does it solve? Connecting a Python repository to Beaker's hosted optimization platform requires wiring an Integration contract, real labeled data, model routing, credentials, and GitHub App access, and mistakes in any step cause failed hosted runs. This Skill guides an agent through the full onboarding loop so the integration validates with a strict smoke check before launch. ## Core Features & Use Cases - Guided onboarding loop: Uses beaker onboarding status as loop control, following its returned next action and relaying developer-owned steps such as GitHub App installation. - Integration contract wiring: Implements Integration, run_setup, run_case, and score_case under .beaker/ while preserving production behavior and adding no tests or CI/CD. - Dataset and credential preflight: Uploads labeled datasets, retains immutable name@revision selectors, declares required_env, and validates with beaker run smoke --strict. - Use Case: A developer asks the agent to set up Beaker in a monorepo service; the Skill discovers the project root, runs beaker init, wires the real model call, pushes a beaker/<timestamp>-<agent> branch, and triggers the first hosted run. ## Quick Start Ask your agent to set up Beaker in this repository and connect it to the real labeled dataset for optimization.
