07-multi-agent

⚠️ This is a reference solution. Workshop participants should follow the steps in the root README and only consult this if stuck.

Solution 07 — TravelBuddy as a runtime multi-agent handoff

Builds on Step 6: keeps every carried capability (function tools, Foundry Toolbox, RAG, the local travel-guide Skill, and the required Foundry response-guardrails Skill) and splits TravelBuddy into a Coordinator plus Flights / Hotels / Activities specialists wired with HandoffBuilder. The Coordinator routes each turn to the specialist that owns it; specialists hand back for synthesis. The whole graph is exposed as one agent via workflow.as_agent(), so hosting is unchanged.

Layout

• travel_assistant/ — the agent code. coordinator.py builds the handoff graph and each specialist gets a sliced capability set; the Coordinator downloads the Foundry skill at runtime into a writable temp dir (<tempdir>/foundry_downloaded_skills/) and serves it plus the local skill via one SkillsProvider. agents/{flights,hotels,activities}/ hold per-specialist agent.yaml + agent.manifest.yaml slices that document each role's tool/RAG/skill boundary. Snapshotted by azd ai agent init.
• travel_indexer/ — the out-of-band Search indexer (provision_index.py, data/destinations.json), a sibling of travel_assistant/ (from Step 5).
• foundry_skills/ — the out-of-band Foundry-skill authoring + upload (provision_skills.py, skills/response-guardrails/SKILL.md), a sibling of travel_assistant/ (from Step 6). Never deployed — azd ai agent init snapshots only travel_assistant/.
• travel_toolbox/ — the toolbox definition (toolbox.yaml).
• .env.example — shared configuration, including FOUNDRY_SKILL_NAMES.

Capability slices

Specialist	Tools	RAG	Skill
Coordinator	—	—	travel-guide, response-guardrails
Flights	get_weather, get_local_time, convert_currency, toolbox (flight fares)	—	—
Hotels	convert_currency, toolbox (web)	destinations index	—
Activities	toolbox (web/reference)	destinations index	—

Run it

1. Copy .env.example to .env and fill in values (including FOUNDRY_SKILL_NAMES).
2. Sign in with Entra ID: az login. Uploading and downloading Foundry skills needs Foundry User (formerly Azure AI User) on the Foundry project, and the account must allow public network access (the Skills API does not support private networking).
3. Install dependencies: pip install -r travel_assistant/requirements.txt.
4. Provision the destinations index once (from Step 5): python travel_indexer/provision_index.py.
5. Upload the Foundry skill once (re-run after editing its SKILL.md): python foundry_skills/provision_skills.py.
6. Start the hosted Responses server: cd travel_assistant && python main.py.

Try: Help me plan a 5-day Tokyo trip: flights from Lisbon, a hotel near Shibuya under €200/night, and a day-trip suggestion. — expect handoffs across all three specialists.

Back to workshop: ../../README.md

TODOs

None.

Upstream attribution

No single upstream sample; the handoff pattern follows the Agent Framework multi-agent handoff orchestration docs.