Name: deploying-scalable-agents
Rating: 74.4 (1 reviews)
Author: microsoft

deploying-scalable-agents

Take a working agent prototype to a scalable, observable production deployment on Microsoft Foundry. Covers deployment patterns (client-hosted, hosted agents, agent workflows), the agent lifecycle, model routing, response caching, evaluation gates, human-in-the-loop approval, observability with OpenTelemetry, cost optimisation, and smoke-testing deployed agents with the AI Smoke Test action. Based on Lesson 16 of AI Agents for Beginners. USE FOR: deploy an agent to production, scale an agent, Microsoft Foundry hosted agent, Foundry Agent Service, model routing, response caching, evaluation gate, release gate, human approval workflow, agent observability, agent tracing, agent cost optimisation, smoke test a hosted agent, production customer support agent. DO NOT USE FOR: building your first agent (start with Lesson 01), running agents locally on-device (use local-ai-agents / Lesson 17), Azure infrastructure provisioning unrelated to agents, non-Foundry deployment targets.

Quality

92%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

85%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, concise overview with a clear gated lifecycle and clean one-level references to the lesson materials. Its only weakness is actionability: it points to the external notebook for runnable code rather than embedding copy-paste-ready examples.

Suggestions

Inline one small copy-paste-ready snippet for the most common task (e.g., a minimal evaluation-gate check or a tracer-wrapped request handler) so the skill is actionable without opening the notebook.

Add a short concrete example of the smoke-test POST request with expected assertion behavior, since the endpoint and role are already named but the call shape is not shown.

Show a minimal model-routing decision snippet (the DIY classifier or Foundry Model Router call) to back the 'route by complexity' lever with executable guidance.

Dimension	Reasoning	Score
Conciseness	Lean and table-driven; it does not explain concepts Claude already knows (no "what is OpenTelemetry/RBAC" prose) and every section earns its tokens — the prototype→production table and prioritized cost levers pack dense guidance into few lines.	3 / 3
Actionability	It names concrete APIs and specifics ("@tool(approval_mode=\"always_require\")", "tracer.start_as_current_span(...)" with attributes "routed.model", "customer.id", the smoke-test endpoint and the Azure AI User role) but explicitly defers executable code to the external notebook, so guidance is specific yet not copy-paste-ready in-body.	2 / 3
Workflow Clarity	The lifecycle is sequenced ("create → version → evaluate (gate) → deploy hosted → observe online → collect failures → repeat") with explicit validation gates ("a version does not ship unless it clears the threshold") and a layered gate model plus a collect-failures→repeat feedback loop.	3 / 3
Progressive Disclosure	SKILL.md acts as an overview pointing to clearly signaled, one-level-deep references (the lesson notebook, README, smoke-test workflow, and tests/ catalog) via markdown links, with no nested reference chains and content appropriately split into themed sections.	3 / 3
	Total	11 / 12 Passed

Description

100%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, specific description that names concrete capabilities, surfaces natural trigger terms, explicitly marks when to use it, and draws clear boundaries against neighboring skills. It uses third person and stays concise without padding.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — "deployment patterns (client-hosted, hosted agents, agent workflows)", "model routing", "response caching", "evaluation gates", "human-in-the-loop approval", "observability with OpenTelemetry", "cost optimisation", "smoke-testing" — matching the multiple-specific-actions anchor.	3 / 3
Completeness	Explicitly answers both what ("Covers deployment patterns... the agent lifecycle, model routing...") and when ("USE FOR: deploy an agent to production..."), plus a DO NOT USE FOR boundary, satisfying the what-AND-when-with-explicit-triggers anchor.	3 / 3
Trigger Term Quality	The USE FOR clause gives natural phrasings a user would say — "deploy an agent to production", "scale an agent", "model routing", "agent observability", "agent tracing", "smoke test a hosted agent", "production customer support agent" — with good coverage and common variations.	3 / 3
Distinctiveness Conflict Risk	A clear Foundry-specific niche with distinct triggers and an explicit DO NOT USE FOR list ("building your first agent", "running agents locally on-device", "Azure infrastructure provisioning unrelated to agents") that prevents overlap with sibling skills.	3 / 3
	Total	12 / 12 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
relative_links	Relative link issues: 4 suspicious	Warning

	Total	15 / 16 Passed

Repository: microsoft/ai-agents-for-beginners
Path: .agents/skills/deploying-scalable-agents/SKILL.md
Commit: 265bf47

Reviewed: about 10 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.