CtrlK
BlogDocsLog inGet started
Tessl Logo

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.

72

Quality

89%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

78%Weight 40%Scale 1-5

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

The content is a well-structured, action-dense overview that names real APIs and roles and sequences the lifecycle with explicit gates. Its main weakness is minor verbosity and pattern-level rather than fully copy-paste-ready code examples.

Suggestions

Trim explanatory asides (e.g. 'a green deploy can still be silent', 'Less control, far less operational surface') to tighten the conciseness score toward 5.

Add one complete, copy-paste-ready code snippet for the request handler (cache → route → trace → run) so the common case is fully executable rather than pattern-level.

Make the evaluation-gate failure path explicit (e.g. 'If pass_rate < threshold: do not deploy; log failures to the offline test set and repeat') to add a validate→fix→retry feedback loop.

DimensionReasoningScore

Conciseness

The body is dense and largely assumes Claude's competence (no explanations of what Foundry/OTel/RBAC are), but a few explanatory asides ('a green deploy can still be silent', 'Less control, far less operational surface') could be trimmed, landing just below the lean-and-efficient anchor.

4 / 5

Actionability

Provides concrete, named mechanisms — FoundryChatClient(...)+provider.as_agent(...), @tool(approval_mode="always_require"), tracer.start_as_current_span(...) with routed.model/customer.id attributes, the exact POST endpoint path, the Azure AI User role, and the token audience — but most guidance is pattern-level rather than a full copy-paste-ready code block.

4 / 5

Workflow Clarity

An explicit lifecycle sequence with a ship/no-ship evaluation gate and a layered smoke-test → offline-eval → online-eval gate structure gives clear sequencing and validation, but an explicit validate→fix→retry recovery loop is implied rather than spelled out.

4 / 5

Progressive Disclosure

No bundle files exist; the body is organized into clean headed sections and points only one-level-deep to external course materials (lesson README, notebook, workflow file, tests catalog) via clearly signaled relative links, with no large reference dump inlined.

5 / 5

Total

17

/

20

Passed

Description

100%Weight 40%Scale 1-5

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

The description is excellent: it states concrete capabilities, supplies comprehensive natural trigger terms, explicitly answers both what and when, and carves out a distinct Foundry-specific niche with negative-boundary guidance. No changes needed.

DimensionReasoningScore

Specificity

Lists many concrete actions — deployment patterns, agent lifecycle, model routing, response caching, evaluation gates, human-in-the-loop approval, OpenTelemetry observability, cost optimisation, and smoke-testing — giving comprehensive rather than merely 'several' specific actions.

5 / 5

Completeness

Explicitly answers 'what' (the opening sentence plus the 'Covers...' list) and 'when' (the 'USE FOR' trigger phrases), with a 'DO NOT USE FOR' boundary clause for extra clarity.

5 / 5

Trigger Term Quality

The 'USE FOR' clause gives comprehensive natural terms with synonyms (evaluation gate/release gate, observability/tracing, Foundry Agent Service/hosted agent), matching the comprehensive synonym-and-variation anchor.

5 / 5

Distinctiveness Conflict Risk

A clear Microsoft Foundry / hosted-agent niche with distinct triggers and a 'DO NOT USE FOR' carve-out (Lesson 01, Lesson 17, non-Foundry targets), giving minimal conflict risk.

5 / 5

Total

20

/

20

Passed

Validation

93%

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

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

relative_links

Relative link issues: 4 suspicious

Warning

Total

15

/

16

Passed

Repository
microsoft/ai-agents-for-beginners
Reviewed

Table of Contents

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.