v0.2.0

Agents Primer 0.2.0 expands the repository from workflow sketches into executable OpenAI Agents SDK lessons.

Highlights:

• SDK-backed examples for typed context, turn limits, retries, guardrails, approvals, resumable RunState, tracing, MCP, tool access, hosted code execution, and typed evaluations.
• Concise orchestration lessons for fan-out, evaluation/refinement, plan execution, manager/worker delegation, partial failure, dependency scheduling, and replanning.
• Offline ScriptedModel coverage that exercises the real Runner and SDK control flow without API calls.
• live-core, live-orchestration, and live-all Make targets for API-backed smoke runs.
• 39 tests covering all 45 numbered examples.

See the full CHANGELOG for details.

v0.1.0

Changed

• Prepared the repository for public release with a GitHub-first badge policy and a strict static-type-checking quality gate.
• Pinned CI to Python 3.12 with locked dependency synchronization.