Azure Functions hosted skills: Queue starter
The smallest useful hosted skill that reacts to an Azure Storage queue. Drop a support message on the support-requests queue and the skill triages it — category, urgency, sentiment, a one-line summary, and a suggested first reply. There are no custom tools: the whole skill is one Markdown file.
Use this as a starting point, the same way functions-daily-blog-hosted-skills-azd is the starting point for timer-triggered skills.
support message ──► Storage queue 'support-requests' ──► support_triage.agent.md ──► logs (Application Insights)
(JSON) hosted skill (Flex Consumption) └─► email via Office 365 (optional)
└─ code interpreter (ACA dynamic sessions)
What gets deployed
azd up creates a resource group with:
- A Flex Consumption function app (Python 3.13) with a user-assigned managed identity.
- A storage account (no shared key access) for the app and the
support-requestsqueue. Messages that fail five times go tosupport-requests-poison. - A Microsoft Foundry project with a
gpt-5.4deployment. - An Azure Container Apps dynamic session pool, so the skill can run Python when it needs to (calculations, formatting, dates).
- Optional: a Connector Namespace with an Office 365 Outlook connection, only when
TO_EMAILis set. - Application Insights and Log Analytics.
All connections use the managed identity. Your signed-in account also gets data-plane roles so you can send messages and run the app locally.
Prerequisites
- An Azure subscription with quota for
gpt-5.4(GlobalStandard) in your region. - Azure Developer CLI (
azd) and Azure CLI (az). - Python 3.13.
- For local runs: Azure Functions Core Tools 4.x and Azurite.
Deploy to Azure
cd starter-queue
azd auth login
azd env new starter-queue
azd env set TO_EMAIL you@example.com # optional: email each triage result
azd up
If you set TO_EMAIL, authorize the Office 365 connection once: open the Connector Namespace (cg-...) in the Azure portal, select the office365 connection, and sign in. Without TO_EMAIL, no email is sent.
By default, the hosted-skills runtime does not include the full model response in traces. To inspect sample JSON in Application Insights, add ENABLE_SENSITIVE_DATA=true to the Function App's application settings and restart the app. This exposes full model responses in logs; use it only with sample or otherwise non-sensitive data, since responses may contain personal data and are subject to your log access and retention settings.
Try it
python -m venv .venv
# Windows: .venv\Scripts\activate macOS/Linux: source .venv/bin/activate
pip install --pre -r requirements-dev.txt
python scripts/demo.py send # send both samples in samples/
python scripts/demo.py send --sample T-1001
Each message starts one run. The skill ends with the triage as JSON, for example:
{"ticket_id": "T-1001", "category": "billing", "urgency": "high", "sentiment": "negative",
"summary": "Customer was charged twice for order 58213 and needs the duplicate refunded before Friday.",
"suggested_reply": "Hi Maria, I'm sorry about the double charge on order 58213..."}
Find invocation details in Application Insights (traces for the support_triage operation), or the result in your inbox when TO_EMAIL is set. You can also put any JSON or plain-text message on the queue from the Azure portal.
python scripts/demo.py e2e sends unique copies and runs a best-effort smoke check for successful invocations in Application Insights. It does not correlate invocations with the submitted messages, so it is not proof that each sample was processed; unrelated runs or telemetry delays can affect the result.
Run locally
Run
azd up(orazd provision) once. Local runs still use the deployed Foundry project and session pool.Start Azurite.
Create the settings file and fill in the values from
azd env get-values(FOUNDRY_PROJECT_ENDPOINT,ACA_SESSION_POOL_ENDPOINT, and optionallyTO_EMAIL/O365_MCP_SERVER_URL):cp src/local.settings.json.sample src/local.settings.jsonStart the app and send messages to Azurite from another terminal:
cd src && func start python scripts/demo.py send --local
Locally the app uses your az login identity.
Make it yours
| To change | Edit |
|---|---|
| What the skill does with each message | the Markdown body of src/support_triage.agent.md |
| Queue name or connection | trigger.args in src/support_triage.agent.md, plus queueName in infra/main.bicep |
| Retries before poison queue, batch size | extensions.queues in src/host.json |
| Model or timeout | src/agents.config.yaml, and azd env set FOUNDRY_MODEL ... |
| Add tools | Add Python tools under src/tools/ or remote MCP servers to src/mcp.json |
Files
starter-queue/
├── azure.yaml
├── infra/ Bicep: function app, storage + queue, Foundry, session pool, optional connector
├── samples/ sample support messages
├── scripts/demo.py send samples, run the e2e check
└── src/
├── support_triage.agent.md the hosted skill (trigger + instructions)
├── agents.config.yaml model, timeout, code interpreter
├── mcp.json optional Office 365 Outlook MCP server
├── function_app.py create_function_app()
└── host.json