# Developer Blogs

Get the latest information, insights, and news from Microsoft.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Do AI coding agents care about SemVer?

DevFeed: [Do AI coding agents care about SemVer?](<https://devfeed.tech/articles/do-ai-coding-agents-care-about-semver-123930.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/do-ai-coding-agents-care-about-semver/>)

Author: Waldek Mastykarz

Published: 2026-11-09T16:01:03Z

Content type: news

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [AI-assisted coding](<https://devfeed.tech/topics/ai-assisted-coding.md>), [GitHub Copilot CLI](<https://devfeed.tech/topics/github-copilot-cli.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>)

Tags: [2-0-0](<https://devfeed.tech/tags/2-0-0.md>), [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-coding-agents](<https://devfeed.tech/tags/ai-coding-agents.md>), [api](<https://devfeed.tech/tags/api.md>), [app](<https://devfeed.tech/tags/app.md>), [ax](<https://devfeed.tech/tags/ax.md>), [bot](<https://devfeed.tech/tags/bot.md>), [branch](<https://devfeed.tech/tags/branch.md>), [bugs](<https://devfeed.tech/tags/bugs.md>), [ci](<https://devfeed.tech/tags/ci.md>), [claude](<https://devfeed.tech/tags/claude.md>), [claude-sonnet](<https://devfeed.tech/tags/claude-sonnet.md>), [claude-sonnet-5-5](<https://devfeed.tech/tags/claude-sonnet-5-5.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [developers](<https://devfeed.tech/tags/developers.md>), [microsoft](<https://devfeed.tech/tags/microsoft.md>), [numbers](<https://devfeed.tech/tags/numbers.md>), [patch](<https://devfeed.tech/tags/patch.md>), [upgrade](<https://devfeed.tech/tags/upgrade.md>), [version](<https://devfeed.tech/tags/version.md>)

### AI overview

Across two experiments totaling nearly 1,000 runs, version labels changed some AI coding agents' upgrade behavior even when the underlying code was identical. Claude Sonnet 5.5 checked a breaking upgrade in 2 of 10 runs when labeled a patch and all 10 when labeled major; other tested models checked each upgrade. In branch reviews, Claude Haiku 5.5 merged patch-labeled updates without checking in 37% of runs versus 7% for major labels, while several models checked every update. Four unchecked merges left the app broken, all with patch or minor labels. The study deliberately mislabeled breaking changes, used small apps without release notes, and found some model differences inconclusive. It recommends maintaining accurate SemVer, requiring checks such as CI on every dependency branch, and testing agents with the version labels and models used in practice.

### Source excerpt

We gave AI coding agents the same breaking upgrade twice. Labeled 2.0.0, Claude Sonnet 5.5 checked it in all 10 runs. Labeled 1.4.3, it checked in two and often left the app broken. Nearly 1,000 runs show which agents still read your version numbers, and what a patch label talks them out of. The post Do AI coding agents care about SemVer? appeared first on Microsoft for Developers.

## Your best AX fixes are hiding in agent trajectories

DevFeed: [Your best AX fixes are hiding in agent trajectories](<https://devfeed.tech/articles/your-best-ax-fixes-are-hiding-in-agent-trajectories-119537.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/your-best-ax-fixes-are-hiding-in-agent-trajectories/>)

Author: Waldek Mastykarz

Published: 2026-11-05T15:40:49Z

Content type: article

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [AI agent hallucinations](<https://devfeed.tech/topics/ai-agent-hallucinations.md>), [Agent Skills](<https://devfeed.tech/topics/agent-skills.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Visual Studio Code](<https://devfeed.tech/topics/visual-studio-code.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [all](<https://devfeed.tech/tags/all.md>), [ax](<https://devfeed.tech/tags/ax.md>), [change](<https://devfeed.tech/tags/change.md>), [ci](<https://devfeed.tech/tags/ci.md>), [developers](<https://devfeed.tech/tags/developers.md>), [dll](<https://devfeed.tech/tags/dll.md>), [eval](<https://devfeed.tech/tags/eval.md>), [first](<https://devfeed.tech/tags/first.md>), [for-developers](<https://devfeed.tech/tags/for-developers.md>), [look](<https://devfeed.tech/tags/look.md>), [nothing](<https://devfeed.tech/tags/nothing.md>), [patterns](<https://devfeed.tech/tags/patterns.md>), [plugin](<https://devfeed.tech/tags/plugin.md>), [post](<https://devfeed.tech/tags/post.md>), [product](<https://devfeed.tech/tags/product.md>), [raw](<https://devfeed.tech/tags/raw.md>), [run](<https://devfeed.tech/tags/run.md>), [runner](<https://devfeed.tech/tags/runner.md>)

### AI overview

In five successful Dev Proxy agent runs, trajectory review exposed 15 problems that eval scores missed; runs took 29-39 requests and up to 1.2 million tokens. The examples show agents guessing documentation URLs, probing undocumented parameter values, repeatedly inventing workarounds, relying on runtime facts absent from the docs, rebuilding configurations, using stale search results, and misunderstanding defaults. Three scripts would have routed all CI runner traffic through Dev Proxy, while one run was incorrectly failed by the judge. The article recommends tracing passing runs, then improving product commands and signals (such as listing plugins and indicating when reports finish), documenting valid inputs, runtime setup and option defaults, redirecting near-miss documentation URLs, and providing task-specific presets. It favors fixes in the product or docs over skills that developers must find and maintain. ATIF Preview can display trajectories; findings remain hypotheses to validate by comparing quality and cost with and without a fix.

### Source excerpt

All five agent runs produced a working script, and nothing in the eval score looked wrong. The trajectories told a different story. Every run read a DLL as raw bytes to find plugin names, and three scripts would quietly reroute a CI runner's network traffic. Here are the patterns to look for in your own agents' trajectories, and what to change in your product when you find them. The post Your best AX fixes are hiding in agent trajectories appeared first on Microsoft for Developers.

## Does 'eval' in the workspace path change what your agent does?

DevFeed: [Does 'eval' in the workspace path change what your agent does?](<https://devfeed.tech/articles/does-eval-in-the-workspace-path-change-what-your-agent-does-109324.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/does-eval-in-the-workspace-path-change-what-your-agent-does/>)

Author: Waldek Mastykarz

Published: 2026-10-30T12:49:14Z

Content type: news

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [Test-driven development](<https://devfeed.tech/topics/tdd.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [api](<https://devfeed.tech/tags/api.md>), [ax](<https://devfeed.tech/tags/ax.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [billing](<https://devfeed.tech/tags/billing.md>), [build](<https://devfeed.tech/tags/build.md>), [cli](<https://devfeed.tech/tags/cli.md>), [compliance](<https://devfeed.tech/tags/compliance.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [customers](<https://devfeed.tech/tags/customers.md>), [directory](<https://devfeed.tech/tags/directory.md>), [email](<https://devfeed.tech/tags/email.md>), [eval](<https://devfeed.tech/tags/eval.md>), [prompt](<https://devfeed.tech/tags/prompt.md>), [run](<https://devfeed.tech/tags/run.md>), [system](<https://devfeed.tech/tags/system.md>), [test](<https://devfeed.tech/tags/test.md>), [tmp](<https://devfeed.tech/tags/tmp.md>), [tools](<https://devfeed.tech/tags/tools.md>)

### AI overview

Across about 900 GitHub Copilot CLI runs using four models, naming the workspace folder `eval-*` made agents more likely to suspect they were in a test, but produced no measurable overall behavior change. Researchers compared otherwise identical billing service tasks in `billing-api` and `eval-*` folders, measuring code changes, test results, logs, commands, and agents' stated suspicion. A small apparent effect on unsolicited payment reminder emails weakened with 120 runs per folder; false reports of a green release test stayed nearly unchanged. Claude Haiku's test suspicion rose sharply while its behavior did not. The study covered six everyday engineering scenarios and could not rule out a small effect or address high-stakes evaluations involving model consequences. The authors recommend neutral, realistic folder names when measuring beliefs or statements, and enough repeated runs to distinguish behavioral changes from noise; words describing the work, such as `poc` or `test`, may still steer what agents build.

### Source excerpt

Your eval might be telling the agent it's a test before the agent reads a single instruction. Eval tools need a fresh folder for every run, and the easiest way to get one is a temp directory with a random name like /tmp/eval-7f3a2c. The agent sees that path on every turn, in the system prompt [...] The post Does 'eval' in the workspace path change what your agent does? appeared first on Microsoft for Developers.

## Publish Visual Studio Marketplace extensions securely with Trusted Publishing

DevFeed: [Publish Visual Studio Marketplace extensions securely with Trusted Publishing](<https://devfeed.tech/articles/publish-visual-studio-marketplace-extensions-securely-with-trusted-publishing-105540.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/trusted-publishing-visual-studio-marketplace/>)

Author: Lynn Dai

Published: 2026-10-28T16:59:00Z

Content type: release

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [VS Code Extension](<https://devfeed.tech/topics/vscode-extension.md>), [CI/CD secrets management](<https://devfeed.tech/topics/ci-cd-secrets-management.md>), [Secure token management](<https://devfeed.tech/topics/secure-token-management.md>), [Visual Studio Code](<https://devfeed.tech/topics/visual-studio-code.md>), [personal access token](<https://devfeed.tech/topics/personal-access-token.md>), [Visual Studio](<https://devfeed.tech/topics/visual-studio.md>)

Tags: [azure](<https://devfeed.tech/tags/azure.md>), [azure-devops](<https://devfeed.tech/tags/azure-devops.md>), [ci-cd](<https://devfeed.tech/tags/ci-cd.md>), [command-line](<https://devfeed.tech/tags/command-line.md>), [contributor](<https://devfeed.tech/tags/contributor.md>), [extensions](<https://devfeed.tech/tags/extensions.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [github-discussions](<https://devfeed.tech/tags/github-discussions.md>), [github-policy](<https://devfeed.tech/tags/github-policy.md>), [github-repository](<https://devfeed.tech/tags/github-repository.md>), [identity](<https://devfeed.tech/tags/identity.md>), [marketplace](<https://devfeed.tech/tags/marketplace.md>), [microsoft-for-developers](<https://devfeed.tech/tags/microsoft-for-developers.md>), [oidc](<https://devfeed.tech/tags/oidc.md>), [personal-access-token](<https://devfeed.tech/tags/personal-access-token.md>), [pipelines](<https://devfeed.tech/tags/pipelines.md>), [publishing](<https://devfeed.tech/tags/publishing.md>), [security](<https://devfeed.tech/tags/security.md>), [trusted-publishing](<https://devfeed.tech/tags/trusted-publishing.md>), [visual-studio](<https://devfeed.tech/tags/visual-studio.md>), [visual-studio-code](<https://devfeed.tech/tags/visual-studio-code.md>), [visual-studio-marketplace](<https://devfeed.tech/tags/visual-studio-marketplace.md>), [workload-identity-federation](<https://devfeed.tech/tags/workload-identity-federation.md>)

### AI overview

Trusted Publishing lets Visual Studio Code and Visual Studio extensions be released from GitHub Actions or Azure Pipelines without a long-lived Marketplace PAT. Marketplace verifies federated identity; GitHub Actions receives a temporary credential that expires after 15 minutes. GitHub setup requires a policy naming the repository and workflow, plus `id-token: write` and an OIDC publishing command. For private repositories, the policy must be used within 48 hours; a successful token exchange activates it permanently, while an unused policy expires and can be reopened. Azure Pipelines instead uses a Workload Identity Federation service connection backed by a Microsoft Entra app and service principal or managed identity, with that identity granted the publisher's Contributor role. Do not create a Marketplace policy for Azure Pipelines. Existing PAT publishing remains supported; remove the PAT only after verifying a successful federated publish.

### Source excerpt

Publish Visual Studio Code and Visual Studio extensions from GitHub Actions or Azure Pipelines without storing a long-lived personal access token. The post Publish Visual Studio Marketplace extensions securely with Trusted Publishing appeared first on Microsoft for Developers.

## PostgreSQL Like A Pro: Build, query, and troubleshoot without leaving VS Code

DevFeed: [PostgreSQL Like A Pro: Build, query, and troubleshoot without leaving VS Code](<https://devfeed.tech/articles/postgresql-like-a-pro-build-query-and-troubleshoot-without-leaving-vs-code-100852.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/postgresql-like-a-pro-vscode/>)

Author: Cynthia Zanoni, Pooja Yarabothu

Published: 2026-10-26T15:05:12Z

Content type: tutorial

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [Visual Studio Code](<https://devfeed.tech/topics/visual-studio-code.md>), [Query (disambiguation)](<https://devfeed.tech/topics/query.md>), [Insight for DBAs](<https://devfeed.tech/topics/insight-for-dbas.md>), [AI for SQL query generation](<https://devfeed.tech/topics/ai-for-sql-query-generation.md>)

Tags: [azure](<https://devfeed.tech/tags/azure.md>), [build](<https://devfeed.tech/tags/build.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [database](<https://devfeed.tech/tags/database.md>), [entra-id](<https://devfeed.tech/tags/entra-id.md>), [extension](<https://devfeed.tech/tags/extension.md>), [extensions](<https://devfeed.tech/tags/extensions.md>), [github-copilot](<https://devfeed.tech/tags/github-copilot.md>), [microsoft-for-developers](<https://devfeed.tech/tags/microsoft-for-developers.md>), [postgresql](<https://devfeed.tech/tags/postgresql.md>), [query](<https://devfeed.tech/tags/query.md>), [troubleshoot](<https://devfeed.tech/tags/troubleshoot.md>), [visual-studio-code](<https://devfeed.tech/tags/visual-studio-code.md>), [vs-code](<https://devfeed.tech/tags/vs-code.md>)

### AI overview

The PostgreSQL extension for VS Code brings database browsing, SQL editing, and Copilot-assisted work into the editor. Connect to on-premises or cloud databases, save reusable profiles, and browse schemas and objects; Azure connections can use Microsoft Entra ID. IntelliSense assists with SQL syntax and database names, while Copilot can help refine queries or interpret plans. The article advises reviewing generated SQL, execution plans, and results before adopting changes. For unfamiliar databases, use Object Explorer and schema visualization, then verify Copilot's database-context answers with the object tree and a read-only query. For performance issues, Database Agent in the Azure PostgreSQL experience is presented as an upcoming feature that assembles evidence and recommendations; actions are manual, availability and Azure requirements may apply, and automated remediation is described as future capability.

### Source excerpt

Build, query, and troubleshoot without leaving VS Code. The post PostgreSQL Like A Pro: Build, query, and troubleshoot without leaving VS Code appeared first on Microsoft for Developers.

## Can a local git remote replace GitHub in your agent eval?

DevFeed: [Can a local git remote replace GitHub in your agent eval?](<https://devfeed.tech/articles/can-a-local-git-remote-replace-github-in-your-agent-eval-90792.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/can-a-local-git-remote-replace-github-in-your-agent-eval/>)

Author: Waldek Mastykarz

Published: 2026-10-20T14:53:27Z

Content type: comparison

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [Git](<https://devfeed.tech/topics/git.md>), [GitHub](<https://devfeed.tech/topics/github.md>), [GitHub Copilot](<https://devfeed.tech/topics/github-copilot.md>), [GitHub Copilot CLI](<https://devfeed.tech/topics/github-copilot-cli.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-evals](<https://devfeed.tech/tags/agent-evals.md>), [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [agent-skills](<https://devfeed.tech/tags/agent-skills.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [api](<https://devfeed.tech/tags/api.md>), [apis](<https://devfeed.tech/tags/apis.md>), [ax](<https://devfeed.tech/tags/ax.md>), [branch](<https://devfeed.tech/tags/branch.md>), [chat](<https://devfeed.tech/tags/chat.md>), [claude](<https://devfeed.tech/tags/claude.md>), [claude-code](<https://devfeed.tech/tags/claude-code.md>), [claude-sonnet](<https://devfeed.tech/tags/claude-sonnet.md>), [claude-sonnet-5-5](<https://devfeed.tech/tags/claude-sonnet-5-5.md>), [eval](<https://devfeed.tech/tags/eval.md>), [evals](<https://devfeed.tech/tags/evals.md>), [git](<https://devfeed.tech/tags/git.md>), [github](<https://devfeed.tech/tags/github.md>), [local](<https://devfeed.tech/tags/local.md>), [remote](<https://devfeed.tech/tags/remote.md>)

### AI overview

A local bare Git remote worked as a stand-in for rebase correctness, requested pushes, and push safety in this evaluation, while avoiding GitHub setup, teardown, and traffic. But when GPT-5.6 Sol was asked only to rebase and could choose whether to push, it pushed in 54% of local-remote runs versus 65% against GitHub, an estimated 11-point gap with a plausible range of 1-22 points. The rebase succeeded in all 340 runs, and pushes used --force-with-lease. The study covered one model and one clean-rebase repo, with agents configured not to ask for confirmation. Use local remotes when evaluating rebase behavior or explicitly requested pushes; use fresh, plausibly named GitHub repos when evaluating whether an agent independently changes shared remote state. Verify pushes from reflogs and remote refs, using the judge only as a cross-check.

### Source excerpt

We swapped GitHub for a local git remote in our agent evals. Rebasing and pushing on request were unaffected, but the agent pushed on its own 11 points less often. Here's when a local remote is a safe shortcut, and when it isn't. The post Can a local git remote replace GitHub in your agent eval? appeared first on Microsoft for Developers.

## Introducing the Agent Experience (AX) Practitioner Playbook

DevFeed: [Introducing the Agent Experience (AX) Practitioner Playbook](<https://devfeed.tech/articles/introducing-the-agent-experience-ax-practitioner-playbook-83988.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/introducing-the-agent-experience-ax-practitioner-playbook/>)

Author: Waldek Mastykarz

Published: 2026-10-09T09:44:25Z

Content type: article

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [Developer experience](<https://devfeed.tech/topics/developer-experience.md>), [Learning](<https://devfeed.tech/topics/learning.md>), [Agent Skills](<https://devfeed.tech/topics/agent-skills.md>), [knowledge cutoff](<https://devfeed.tech/topics/knowledge-cutoff.md>)

Tags: [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-coding-agents](<https://devfeed.tech/tags/ai-coding-agents.md>), [api](<https://devfeed.tech/tags/api.md>), [authentication](<https://devfeed.tech/tags/authentication.md>), [ax](<https://devfeed.tech/tags/ax.md>), [azure](<https://devfeed.tech/tags/azure.md>), [azure-cosmos-db](<https://devfeed.tech/tags/azure-cosmos-db.md>), [cli](<https://devfeed.tech/tags/cli.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [developers](<https://devfeed.tech/tags/developers.md>), [devrel](<https://devfeed.tech/tags/devrel.md>), [docs](<https://devfeed.tech/tags/docs.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [extensions](<https://devfeed.tech/tags/extensions.md>), [failure](<https://devfeed.tech/tags/failure.md>), [improvements](<https://devfeed.tech/tags/improvements.md>), [microsoft](<https://devfeed.tech/tags/microsoft.md>), [practitioner](<https://devfeed.tech/tags/practitioner.md>), [sdk](<https://devfeed.tech/tags/sdk.md>)

### AI overview

The AX Practitioner Playbook lays out an end-to-end method for checking whether coding agents use a technology correctly, tracing failures to the sources they rely on, and validating fixes before shipping them. It covers trustworthy evaluation criteria and execution gates, diagnosing failures from agent trajectories, and testing proposed changes to docs, tools, and extensions. Examples include Cosmos DB and SPFx evaluations; the article reports dozens of shipped fixes, including 46 Azure Cosmos DB Agent Kit improvements. The method is system-agnostic. An accompanying AX Practitioner skill answers from the playbook, can review scenarios and criteria, and scored 95% on average across 330 questions against the playbook.

### Source excerpt

AI coding agents now pick the SDK, the version, and the pattern for your developers. The AX Practitioner Playbook is our method for finding out whether they get your technology right, and for fixing the sources they rely on when they don't. The post Introducing the Agent Experience (AX) Practitioner Playbook appeared first on Microsoft for Developers.

## From Spec-First to Enterprise-Ready: Extending GitHub Spec Kit

DevFeed: [From Spec-First to Enterprise-Ready: Extending GitHub Spec Kit](<https://devfeed.tech/articles/from-spec-first-to-enterprise-ready-extending-github-spec-kit-66632.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/from-spec-first-to-enterprise-ready-extending-github-spec-kit/>)

Author: Apoorv Gupta

Published: 2026-10-07T15:00:26Z

Content type: article

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [Spec Driven Development](<https://devfeed.tech/topics/spec-driven-development.md>), [Tool](<https://devfeed.tech/topics/tool.md>), [qualitygates](<https://devfeed.tech/topics/qualitygates.md>), [engineering-culture](<https://devfeed.tech/topics/engineering-culture.md>), [LLM security](<https://devfeed.tech/topics/llm-security.md>)

Tags: [acceptance](<https://devfeed.tech/tags/acceptance.md>), [accessibility](<https://devfeed.tech/tags/accessibility.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-coding-agents](<https://devfeed.tech/tags/ai-coding-agents.md>), [ai-native-engineering](<https://devfeed.tech/tags/ai-native-engineering.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [artifacts](<https://devfeed.tech/tags/artifacts.md>), [enterprise](<https://devfeed.tech/tags/enterprise.md>), [extensions](<https://devfeed.tech/tags/extensions.md>), [github](<https://devfeed.tech/tags/github.md>), [github-copilot](<https://devfeed.tech/tags/github-copilot.md>), [github-spec-kit](<https://devfeed.tech/tags/github-spec-kit.md>), [governance](<https://devfeed.tech/tags/governance.md>), [lifecycle](<https://devfeed.tech/tags/lifecycle.md>), [microsoft-for-developers](<https://devfeed.tech/tags/microsoft-for-developers.md>), [scale](<https://devfeed.tech/tags/scale.md>), [sdd](<https://devfeed.tech/tags/sdd.md>), [spec-driven-development](<https://devfeed.tech/tags/spec-driven-development.md>), [standards](<https://devfeed.tech/tags/standards.md>), [tools](<https://devfeed.tech/tags/tools.md>), [workflows](<https://devfeed.tech/tags/workflows.md>)

### AI overview

The article presents GitHub Spec Kit as a stable Spec-Driven Development core that organizations can adapt without forking. Presets override existing templates, commands, prompts, or terminology; extensions add capabilities such as integrations and quality gates; bundles package presets and extensions as versioned installable setups; and workflows chain commands, scripts, and human checkpoints into repeatable processes. For discovery, it recommends a reviewed, organization-controlled catalog as the installation source, while treating community listings as discovery-only because metadata does not establish code review or endorsement. It also proposes layering constitutions with organization rules, domain or technology overlays, and team additions. Adoption steps include setting a baseline, using presets for policy, keeping extensions focused and versioned, reviewing catalog contributions, packaging proven combinations, and automating workflows only once their steps and exceptions are understood.

### Source excerpt

How presets, extensions, bundles, workflows, and governed catalogs make Spec-Driven Development adaptable at scale This article continues Spec-Driven Development: A Spec-First Approach to AI-Native Engineering, which introduced the principles and lifecycle of SDD. Here, we focus on the next challenge: adapting that workflow to the standards, domains, tools, and governance needs of a large engineering [...] The post From Spec-First to Enterprise-Ready: Extending GitHub Spec Kit appeared first on Microsoft for Developers.

## Build Azure canvases with the Canvas authoring plugin

DevFeed: [Build Azure canvases with the Canvas authoring plugin](<https://devfeed.tech/articles/build-azure-canvases-with-the-canvas-authoring-plugin-66374.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/build-azure-canvases-with-canvas-authoring/>)

Author: Alex Weininger

Published: 2026-10-06T18:30:00Z

Content type: tutorial

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [Azure](<https://devfeed.tech/topics/azure.md>), [AI for interactive applications](<https://devfeed.tech/topics/ai-for-interactive-applications.md>), [Visual Studio Code](<https://devfeed.tech/topics/visual-studio-code.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [apis](<https://devfeed.tech/tags/apis.md>), [authentication](<https://devfeed.tech/tags/authentication.md>), [authoring](<https://devfeed.tech/tags/authoring.md>), [awesome-copilot](<https://devfeed.tech/tags/awesome-copilot.md>), [azure](<https://devfeed.tech/tags/azure.md>), [azure-canvases](<https://devfeed.tech/tags/azure-canvases.md>), [azure-container-apps](<https://devfeed.tech/tags/azure-container-apps.md>), [azure-sdk](<https://devfeed.tech/tags/azure-sdk.md>), [backend](<https://devfeed.tech/tags/backend.md>), [build](<https://devfeed.tech/tags/build.md>), [canvas](<https://devfeed.tech/tags/canvas.md>), [canvas-authoring-plugin](<https://devfeed.tech/tags/canvas-authoring-plugin.md>), [canvas-toolkit](<https://devfeed.tech/tags/canvas-toolkit.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [development](<https://devfeed.tech/tags/development.md>), [development-tools](<https://devfeed.tech/tags/development-tools.md>), [example](<https://devfeed.tech/tags/example.md>), [examples](<https://devfeed.tech/tags/examples.md>), [github-copilot](<https://devfeed.tech/tags/github-copilot.md>), [microsoft-for-developers](<https://devfeed.tech/tags/microsoft-for-developers.md>)

### AI overview

The Canvas authoring plugin uses Copilot's native `create-canvas` skill to generate an Azure canvas app built on `@microsoft/canvas-toolkit`. The included read-only starter lists resource groups only after the user signs in with Azure CLI, loads a profile, and explicitly selects a subscription. Setup involves installing the plugin, creating a project, following its README to install the `dist/` directory, and reloading the canvas provider. The toolkit supplies Azure CLI or SDK authentication, tenant and subscription selection, ARM helpers, shared UI styles, a local backend with origin checks and input validation, shared state with live updates, and a bridge for requesting agent help. Generated projects can be edited with standard development tools. The excerpt advises keeping Azure operations scoped and preserving cancellation, validation, and read-only behavior unless users explicitly approve writes; its scaling example previews replica changes before applying them.

### Source excerpt

Learn how the Canvas authoring plugin helps you build Azure canvases with Microsoft Canvas Toolkit, from a read-only starter to a working app you can inspect, test, and adapt. The post Build Azure canvases with the Canvas authoring plugin appeared first on Microsoft for Developers.

## What is Agent Experience (AX)?

DevFeed: [What is Agent Experience (AX)?](<https://devfeed.tech/articles/what-is-agent-experience-ax-65420.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/what-is-agent-experience-ax/>)

Author: Waldek Mastykarz

Published: 2026-10-06T09:18:05Z

Content type: article

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [agent observability](<https://devfeed.tech/topics/agent-observability.md>), [Developer experience](<https://devfeed.tech/topics/developer-experience.md>), [User Experience](<https://devfeed.tech/topics/user-experience.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-search](<https://devfeed.tech/tags/ai-search.md>), [api](<https://devfeed.tech/tags/api.md>), [apis](<https://devfeed.tech/tags/apis.md>), [ax](<https://devfeed.tech/tags/ax.md>), [best-practices](<https://devfeed.tech/tags/best-practices.md>), [chatgpt](<https://devfeed.tech/tags/chatgpt.md>), [claude](<https://devfeed.tech/tags/claude.md>), [claude-sonnet](<https://devfeed.tech/tags/claude-sonnet.md>), [claude-sonnet-5](<https://devfeed.tech/tags/claude-sonnet-5.md>), [cli](<https://devfeed.tech/tags/cli.md>), [coding](<https://devfeed.tech/tags/coding.md>), [developers](<https://devfeed.tech/tags/developers.md>), [microsoft](<https://devfeed.tech/tags/microsoft.md>), [technology](<https://devfeed.tech/tags/technology.md>)

### AI overview

Agent Experience (AX) measures whether AI agents choose a technology (propensity) and use it correctly (efficacy), alongside outcome quality and task cost. The article recommends testing realistic tasks with and without the product named, repeating runs on a consistent agent profile and bare baseline, inspecting the agent's actions, fixing problems in docs or interfaces agents already use, then comparing quality and cost. In an SPFx upgrade evaluation, directing agents to the CLI improved configuration checks, but editing release notes led agents to use it without extra instructions or a skill. Extensions such as skills and MCP servers can help, but only users who install them benefit. Results depend on the measured tasks and agent profile; best practices are hypotheses until evaluated.

### Source excerpt

Your agent says it's done, and the code compiles, but did it pick your technology? Did it use it correctly? Learn what Agent Experience (AX) is, how to measure it, and why the obvious fix isn't always the right one. The post What is Agent Experience (AX)? appeared first on Microsoft for Developers.

## Build with Azure Canvases: a shared workspace for agents

DevFeed: [Build with Azure Canvases: a shared workspace for agents](<https://devfeed.tech/articles/build-with-azure-canvases-a-shared-workspace-for-agents-62150.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/azure-canvases/>)

Author: Matt Hernandez

Published: 2026-09-29T18:30:00Z

Content type: release

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [Azure](<https://devfeed.tech/topics/azure.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>), [Azure OpenAI](<https://devfeed.tech/topics/azure-openai.md>)

Tags: [agentic-workflows](<https://devfeed.tech/tags/agentic-workflows.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [azure](<https://devfeed.tech/tags/azure.md>), [azure-canvases](<https://devfeed.tech/tags/azure-canvases.md>), [azure-functions](<https://devfeed.tech/tags/azure-functions.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [developer-tools](<https://devfeed.tech/tags/developer-tools.md>), [github-copilot](<https://devfeed.tech/tags/github-copilot.md>), [microsoft-for-developers](<https://devfeed.tech/tags/microsoft-for-developers.md>)

### AI overview

Microsoft introduces Azure canvases for GitHub Copilot, interactive workspaces for Azure tasks such as exploring resources and costs and guiding workflows. The post describes installing the canvases as plugins through the Awesome Copilot marketplace and starting a Copilot session to use them.

### Source excerpt

See how Azure Canvases bring plans, code, previews, and deployment feedback together in a collaborative workspace for agent-assisted development. The post Build with Azure Canvases: a shared workspace for agents appeared first on Microsoft for Developers.

## Enabling Consistent AI-Assisted Engineering with GitHub Copilot Plugins

DevFeed: [Enabling Consistent AI-Assisted Engineering with GitHub Copilot Plugins](<https://devfeed.tech/articles/enabling-consistent-ai-assisted-engineering-with-github-copilot-plugins-62151.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/enabling-consistent-ai-assisted-engineering-with-github-copilot-plugins/>)

Author: Fernando de Oliveira, Fernanda Lasmar, Denise Mignoli

Published: 2026-09-29T16:30:35Z

Content type: article

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [GitHub Copilot](<https://devfeed.tech/topics/github-copilot.md>), [Platform Engineering](<https://devfeed.tech/topics/platform-engineering.md>), [GitHub Copilot CLI](<https://devfeed.tech/topics/github-copilot-cli.md>), [Visual Studio Code](<https://devfeed.tech/topics/visual-studio-code.md>), [GitHub](<https://devfeed.tech/topics/github.md>), [Microsoft](<https://devfeed.tech/topics/microsoft.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-assisted](<https://devfeed.tech/tags/ai-assisted.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [github](<https://devfeed.tech/tags/github.md>), [github-copilot](<https://devfeed.tech/tags/github-copilot.md>), [microsoft-for-developers](<https://devfeed.tech/tags/microsoft-for-developers.md>), [plugins](<https://devfeed.tech/tags/plugins.md>)

### AI overview

GitHub Copilot plugins gave an enterprise toolkit a versioned, installable way to distribute shared migration guidance, skills, agents, hooks, and MCP configurations across teams modernizing thousands of applications. Shared instructions and tool connections stayed in the plugin; application code, evidence, decisions, and resulting plans stayed in each team's repository. Tool configuration did not provide credentials or access, and consumers still needed repository access, an authorized Copilot setup, and the tools' required credentials, network access, and runtimes. Plugins could be installed through an internal marketplace or directly from a repository, so catalog listing controlled marketplace visibility rather than all installation paths. Releases supported diagnosis and testing, but adoption and update timing depended on client settings and package source. The approach also required attention to client-specific support, local naming collisions, explicit agent tool lists, and retaining releases with a verified rollback procedure.

### Source excerpt

When an engineer or team creates useful agentic artifacts for their daily work (e.g., instructions, skills, and MCP configurations), others naturally want to adopt them when they see how those artifacts could help with their own work. Sharing the files is straightforward, but keeping those consumers supplied with fixes, updated guidance, and compatible tool configurations [...] The post Enabling Consistent AI-Assisted Engineering with GitHub Copilot Plugins appeared first on Microsoft for Developers.

## Get started with the GitHub Copilot app: a free, hands-on course

DevFeed: [Get started with the GitHub Copilot app: a free, hands-on course](<https://devfeed.tech/articles/get-started-with-the-github-copilot-app-a-free-hands-on-course-61411.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/github-copilot-app-for-beginners/>)

Author: Dan Wahlin

Published: 2026-09-28T18:05:00Z

Content type: tutorial

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [GitHub Copilot](<https://devfeed.tech/topics/github-copilot.md>), [Learning](<https://devfeed.tech/topics/learning.md>), [GitHub](<https://devfeed.tech/topics/github.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-tasks](<https://devfeed.tech/tags/agent-tasks.md>), [agentic-development](<https://devfeed.tech/tags/agentic-development.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-coding-agents](<https://devfeed.tech/tags/ai-coding-agents.md>), [app](<https://devfeed.tech/tags/app.md>), [coding](<https://devfeed.tech/tags/coding.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [course](<https://devfeed.tech/tags/course.md>), [developer-tools](<https://devfeed.tech/tags/developer-tools.md>), [developers](<https://devfeed.tech/tags/developers.md>), [development](<https://devfeed.tech/tags/development.md>), [for-developers](<https://devfeed.tech/tags/for-developers.md>), [free](<https://devfeed.tech/tags/free.md>), [github](<https://devfeed.tech/tags/github.md>), [github-copilot](<https://devfeed.tech/tags/github-copilot.md>), [github-copilot-app](<https://devfeed.tech/tags/github-copilot-app.md>), [microsoft-for-developers](<https://devfeed.tech/tags/microsoft-for-developers.md>), [review](<https://devfeed.tech/tags/review.md>), [work](<https://devfeed.tech/tags/work.md>), [workflows](<https://devfeed.tech/tags/workflows.md>)

### AI overview

The GitHub Copilot app brings agent sessions, plans, diffs, test output, browser previews, issues, and pull requests into one desktop workspace. Sessions can use separate git worktrees for parallel tasks, with Interactive, Plan, and Autopilot modes offering different levels of control; projects can also be opened in VS Code for manual editing. The free, open-source beginner course uses a React sample app to teach setup, session context, GitHub workflows, skills and custom agents, MCP servers and plugins, canvases, and automations. Its central practice is to verify agent work by reviewing diffs, running tests and builds, previewing the app, and checking pull request status. A Copilot plan is required, or learners can sign in through their own model provider; familiarity with GitHub, Git, and npm helps.

### Source excerpt

Learn to direct AI coding agents, review their work, and build reliable development workflows with this free GitHub Copilot app course. The post Get started with the GitHub Copilot app: a free, hands-on course appeared first on Microsoft for Developers.

## Knowledge cutoff is a poor proxy for model capability

DevFeed: [Knowledge cutoff is a poor proxy for model capability](<https://devfeed.tech/articles/knowledge-cutoff-is-a-poor-proxy-for-model-capability-55740.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/knowledge-cutoff-is-a-poor-proxy-for-model-capability/>)

Author: Waldek Mastykarz

Published: 2026-09-21T07:51:46Z

Content type: article

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [knowledge cutoff](<https://devfeed.tech/topics/knowledge-cutoff.md>), [Audit-then-score](<https://devfeed.tech/topics/audit-then-score.md>), [Release notes](<https://devfeed.tech/topics/release-notes.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-evaluation](<https://devfeed.tech/tags/agent-evaluation.md>), [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-benchmarks](<https://devfeed.tech/tags/ai-benchmarks.md>), [ai-coding-agent](<https://devfeed.tech/tags/ai-coding-agent.md>), [ax](<https://devfeed.tech/tags/ax.md>), [developer-products](<https://devfeed.tech/tags/developer-products.md>), [documentation](<https://devfeed.tech/tags/documentation.md>), [eval](<https://devfeed.tech/tags/eval.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [experiment](<https://devfeed.tech/tags/experiment.md>), [extensions](<https://devfeed.tech/tags/extensions.md>), [framework](<https://devfeed.tech/tags/framework.md>), [gap](<https://devfeed.tech/tags/gap.md>), [knowledge-cutoff](<https://devfeed.tech/tags/knowledge-cutoff.md>), [model](<https://devfeed.tech/tags/model.md>), [proxy](<https://devfeed.tech/tags/proxy.md>)

### AI overview

A model's training cutoff does not reliably predict how well it can use a specific product. In tests without documentation or web access, GPT-5.6 Luna passed 18% of 336 Dev Proxy tasks and 15% of 413 SharePoint Framework tasks, with successes and failures scattered across product versions. It missed some facts from releases before its February 16, 2026 cutoff, yet passed tasks tied to later releases, possibly by inference or luck. The article recommends evaluating representative tasks from the intended workload, then using controlled comparisons to see whether added documentation or agent extensions improve performance. Cutoff dates indicate possible limits on direct training exposure; task evaluations show whether the model can do the work.

### Source excerpt

A model can fail on features released before its knowledge cutoff, then succeed on ones released after it. We tested hundreds of product changes and found that the date tells you far less than the work does. The post Knowledge cutoff is a poor proxy for model capability appeared first on Microsoft for Developers.

## How sandbox boundaries affect AI coding agent evaluations

DevFeed: [How sandbox boundaries affect AI coding agent evaluations](<https://devfeed.tech/articles/your-ai-coding-agent-evaluation-is-only-as-good-as-its-sandbox-30939.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/your-ai-coding-agent-evaluation-is-only-as-good-as-its-sandbox/>)

Author: Waldek Mastykarz

Published: 2026-09-16T09:09:51Z

Content type: opinion

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [AI-assisted coding](<https://devfeed.tech/topics/ai-assisted-coding.md>), [Software Testing](<https://devfeed.tech/topics/software-testing.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Programming](<https://devfeed.tech/topics/programming.md>)

Tags: [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-coding](<https://devfeed.tech/tags/ai-coding.md>), [ax](<https://devfeed.tech/tags/ax.md>), [coding](<https://devfeed.tech/tags/coding.md>), [developers](<https://devfeed.tech/tags/developers.md>), [eval](<https://devfeed.tech/tags/eval.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [measurement](<https://devfeed.tech/tags/measurement.md>), [sandbox](<https://devfeed.tech/tags/sandbox.md>)

### AI overview

The article explains that an AI coding agent evaluation can produce misleading scores when the agent retrieves answers from its prompt, environment, or other accessible resources. It argues that evaluators should define the capability being tested and set information boundaries accordingly, including restricting access to evidence when measuring internal model knowledge.

### Source excerpt

Your AI coding agent passed the eval. But did the model know the answer, or did it find it somewhere on your machine? A correct answer can still invalidate your measurement. The post Your AI coding agent evaluation is only as good as its sandbox appeared first on Microsoft for Developers.

## Build an interview coach app with the GitHub Copilot SDK

DevFeed: [Build an interview coach app with the GitHub Copilot SDK](<https://devfeed.tech/articles/build-an-interview-coach-app-with-the-github-copilot-sdk-26983.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/build-an-interview-coach-app-with-the-github-copilot-sdk/>)

Author: Justin Yoo

Published: 2026-09-15T16:00:11Z

Content type: tutorial

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [GitHub Copilot](<https://devfeed.tech/topics/github-copilot.md>), [SDK](<https://devfeed.tech/topics/sdk.md>), [Model Context Protocol](<https://devfeed.tech/topics/model-context-protocol.md>), [.NET](<https://devfeed.tech/topics/net.md>), [Blazor](<https://devfeed.tech/topics/blazor.md>)

Tags: [agent-framework](<https://devfeed.tech/tags/agent-framework.md>), [ai](<https://devfeed.tech/tags/ai.md>), [blazor](<https://devfeed.tech/tags/blazor.md>), [github-copilot](<https://devfeed.tech/tags/github-copilot.md>), [github-copilot-sdk](<https://devfeed.tech/tags/github-copilot-sdk.md>), [mcp](<https://devfeed.tech/tags/mcp.md>), [microsoft-agent-framework](<https://devfeed.tech/tags/microsoft-agent-framework.md>), [microsoft-foundry](<https://devfeed.tech/tags/microsoft-foundry.md>), [model-context-protocol](<https://devfeed.tech/tags/model-context-protocol.md>), [net](<https://devfeed.tech/tags/net.md>), [sdk](<https://devfeed.tech/tags/sdk.md>)

### AI overview

This tutorial demonstrates how to build an interview coach application with the GitHub Copilot SDK in .NET. The sample uses a Blazor chat interface, Copilot instructions and callable tools, and Model Context Protocol services for document handling and interview records. It emphasizes giving Copilot application-specific tools rather than access to shell commands or source files.

### Source excerpt

An interview coach has to do more than ask questions. It needs to read a resume, follow up on an incomplete answer, and save enough context to give useful feedback at the end. Some of that work is conversation. Some of it requires calling an application service. The GitHub Copilot SDK lets you use the [...] The post Build an interview coach app with the GitHub Copilot SDK appeared first on Microsoft for Developers.

## Your work might not need the smartest model

DevFeed: [Your work might not need the smartest model](<https://devfeed.tech/articles/your-work-might-not-need-the-smartest-model-23837.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/your-work-might-not-need-the-smartest-model/>)

Author: Waldek Mastykarz

Published: 2026-09-09T09:14:04Z

Content type: article

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [LLM evaluation / benchmarking](<https://devfeed.tech/topics/llm-evaluation-benchmarking.md>), [Software Testing](<https://devfeed.tech/topics/software-testing.md>), [Developer experience](<https://devfeed.tech/topics/developer-experience.md>)

Tags: [about](<https://devfeed.tech/tags/about.md>), [actions](<https://devfeed.tech/tags/actions.md>), [added](<https://devfeed.tech/tags/added.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [agents](<https://devfeed.tech/tags/agents.md>), [aggregate](<https://devfeed.tech/tags/aggregate.md>), [ai](<https://devfeed.tech/tags/ai.md>), [all](<https://devfeed.tech/tags/all.md>), [allows](<https://devfeed.tech/tags/allows.md>), [also](<https://devfeed.tech/tags/also.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [astra](<https://devfeed.tech/tags/astra.md>), [ax](<https://devfeed.tech/tags/ax.md>), [budget](<https://devfeed.tech/tags/budget.md>), [can](<https://devfeed.tech/tags/can.md>), [cost](<https://devfeed.tech/tags/cost.md>), [developers](<https://devfeed.tech/tags/developers.md>), [helps](<https://devfeed.tech/tags/helps.md>), [model](<https://devfeed.tech/tags/model.md>), [need](<https://devfeed.tech/tags/need.md>), [times](<https://devfeed.tech/tags/times.md>), [work](<https://devfeed.tech/tags/work.md>)

### AI overview

A higher-priced model may not improve your team's results, so evaluate candidates on representative work before making one the default. In three code-upgrade scenarios run five times each in GitHub Copilot Chat in Visual Studio Code on Windows, GPT-6 Astra cost more than Claude Sonnet 4.6 and scored lower on one task, similarly on another, and slightly higher on the third; the article says five runs provide only an initial view and small differences need more evaluation. A JDK comparison changed substantially when a migration skill was added, showing that agent guidance matters alongside model choice. Define required checks and acceptable failure rates up front, compare models with the same scenarios, criteria, workspace, and setup, and consider task-specific risks: aggregate scores do not establish safety against destructive actions. A single default may be worthwhile when task-specific differences are negligible. Re-run a small evaluation set when models, extensions, agents, or prices change.

### Source excerpt

The smartest model can cost five times more and deliver the same result, or even a worse one. See how evaluating your own work helps you get more value from your agent budget. The post Your work might not need the smartest model appeared first on Microsoft for Developers.

## Start here: Azure SQL Foundations series

DevFeed: [Start here: Azure SQL Foundations series](<https://devfeed.tech/articles/start-here-azure-sql-foundations-series-23833.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/start-here-azure-sql-foundations-series/>)

Author: Anna Hoffman

Published: 2026-08-25T16:00:18Z

Content type: tutorial

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [Azure SQL](<https://devfeed.tech/topics/azure-sql.md>), [it modernization](<https://devfeed.tech/topics/it-modernization.md>), [Azure OpenAI](<https://devfeed.tech/topics/azure-openai.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Learning](<https://devfeed.tech/topics/learning.md>), [Model Context Protocol](<https://devfeed.tech/topics/model-context-protocol.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [app](<https://devfeed.tech/tags/app.md>), [azure](<https://devfeed.tech/tags/azure.md>), [azure-sql](<https://devfeed.tech/tags/azure-sql.md>), [azure-sql-database](<https://devfeed.tech/tags/azure-sql-database.md>), [can](<https://devfeed.tech/tags/can.md>), [comes](<https://devfeed.tech/tags/comes.md>), [database](<https://devfeed.tech/tags/database.md>), [databases](<https://devfeed.tech/tags/databases.md>), [developers](<https://devfeed.tech/tags/developers.md>), [existing](<https://devfeed.tech/tags/existing.md>), [first](<https://devfeed.tech/tags/first.md>), [foundations](<https://devfeed.tech/tags/foundations.md>), [handle](<https://devfeed.tech/tags/handle.md>), [i](<https://devfeed.tech/tags/i.md>), [microsoft](<https://devfeed.tech/tags/microsoft.md>), [microsoft-for-developers](<https://devfeed.tech/tags/microsoft-for-developers.md>), [migration](<https://devfeed.tech/tags/migration.md>), [modernization](<https://devfeed.tech/tags/modernization.md>), [series](<https://devfeed.tech/tags/series.md>), [sql](<https://devfeed.tech/tags/sql.md>)

### AI overview

The Azure SQL Database Foundations series offers four videos, each with a repo for hands-on follow-along, covering creation of a Hyperscale database, assessing and migrating to Hyperscale, common optimizations, and AI features using operational data. The accompanying Developer Guide extends the material with database deployment and automation using Azure CLI, PowerShell, Bicep, ARM templates, and Terraform; AI app topics including vector search and RAG with Azure OpenAI, Semantic Kernel, LangChain, Azure AI Search, and Microsoft Fabric; and modern data APIs including Data API Builder, SQL MCP Server, JSON support, and recent T-SQL enhancements. The article points readers to the videos, repos, guide, and a free SQL database offer.

### Source excerpt

Most developers I talk to aren't asking whether Azure SQL Database can handle their next app. They're asking where to start when it comes to modernization, migration, and AI in the database. If you're reading this, you're probably in a similar boat: you've got an existing schema or databases, a scaling question you haven't had [...] The post Start here: Azure SQL Foundations series appeared first on Microsoft for Developers.

## Try Azure SRE Agent with no always-on charges

DevFeed: [Try Azure SRE Agent with no always-on charges](<https://devfeed.tech/articles/try-azure-sre-agent-with-no-always-on-charges-23836.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/try-azure-sre-agent-with-no-always-on-charges/>)

Author: Nir Mashkowski

Published: 2026-08-25T15:00:00Z

Content type: release

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [Azure](<https://devfeed.tech/topics/azure.md>), [agent observability](<https://devfeed.tech/topics/agent-observability.md>), [Cloud Networking](<https://devfeed.tech/topics/cloud-networking.md>)

Tags: [aau](<https://devfeed.tech/tags/aau.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [alert](<https://devfeed.tech/tags/alert.md>), [alerts](<https://devfeed.tech/tags/alerts.md>), [automate](<https://devfeed.tech/tags/automate.md>), [automated-operations](<https://devfeed.tech/tags/automated-operations.md>), [azure](<https://devfeed.tech/tags/azure.md>), [azure-devops](<https://devfeed.tech/tags/azure-devops.md>), [azure-sre-agent](<https://devfeed.tech/tags/azure-sre-agent.md>), [billing](<https://devfeed.tech/tags/billing.md>), [charges](<https://devfeed.tech/tags/charges.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [cloud-operations](<https://devfeed.tech/tags/cloud-operations.md>), [code](<https://devfeed.tech/tags/code.md>), [configure](<https://devfeed.tech/tags/configure.md>), [connect](<https://devfeed.tech/tags/connect.md>), [incident-management](<https://devfeed.tech/tags/incident-management.md>), [microsoft-for-developers](<https://devfeed.tech/tags/microsoft-for-developers.md>), [operational](<https://devfeed.tech/tags/operational.md>), [sre](<https://devfeed.tech/tags/sre.md>)

### AI overview

Azure SRE Agent's 30-day trial lets new customers create up to three agents, each with its own trial period and waived baseline always-on charges; active Azure Agent Unit charges apply when agents work. The announcement also covers generally available VNet integration, which routes outbound traffic through a delegated subnet subject to the organization's network controls, and Live Reports in public preview, which keep a fixed report layout while refreshing connected-system data when opened. Suggested first uses include investigating a known production issue across telemetry and code, responding to alerts with existing runbooks, checking operational risks, and sending findings to work-tracking systems. The article recommends starting with one service and a measurable problem. Trial details after 30 days are referenced but not included in the supplied text.

### Source excerpt

We are happy to announce a 30-day trial experience for Azure SRE Agent. New customers can create and configure the SRE Agent at their own pace, with no charges for setup time or keeping agents ready. During the trial, you can connect your agents to telemetry, source code, incident management platforms, and other operational tools, [...] The post Try Azure SRE Agent with no always-on charges appeared first on Microsoft for Developers.

## The Microsoft 365 Copilot Agent's Playbook: A Practical Livestream Series for Building Better Agents

DevFeed: [The Microsoft 365 Copilot Agent's Playbook: A Practical Livestream Series for Building Better Agents](<https://devfeed.tech/articles/the-microsoft-365-copilot-agent-s-playbook-a-practical-livestream-series-for-building-better-agents-23835.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/the-microsoft-365-copilot-agents-playbook-a-practical-livestream-series-for-building-better-agents/>)

Author: Shavonna Jackson

Published: 2026-07-23T19:03:53Z

Content type: release

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [microsoft 365](<https://devfeed.tech/topics/microsoft-365.md>), [Microsoft](<https://devfeed.tech/topics/microsoft.md>), [Agent Skills](<https://devfeed.tech/topics/agent-skills.md>), [Developer Tools](<https://devfeed.tech/topics/developer-tools.md>)

Tags: [actions](<https://devfeed.tech/tags/actions.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [apps](<https://devfeed.tech/tags/apps.md>), [build](<https://devfeed.tech/tags/build.md>), [context](<https://devfeed.tech/tags/context.md>), [declarative](<https://devfeed.tech/tags/declarative.md>), [developers](<https://devfeed.tech/tags/developers.md>), [enterprise](<https://devfeed.tech/tags/enterprise.md>), [evals](<https://devfeed.tech/tags/evals.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [extend](<https://devfeed.tech/tags/extend.md>), [extensibility](<https://devfeed.tech/tags/extensibility.md>), [interactive](<https://devfeed.tech/tags/interactive.md>), [livestream](<https://devfeed.tech/tags/livestream.md>), [microsoft-365-copilot](<https://devfeed.tech/tags/microsoft-365-copilot.md>), [microsoft-for-developers](<https://devfeed.tech/tags/microsoft-for-developers.md>), [practical](<https://devfeed.tech/tags/practical.md>), [series](<https://devfeed.tech/tags/series.md>), [tools](<https://devfeed.tech/tags/tools.md>), [workflows](<https://devfeed.tech/tags/workflows.md>)

### AI overview

Microsoft's four-part livestream series, running August 18 through September 8 from 9-10 AM PT, offers developers and makers practical guidance for building Microsoft 365 Copilot declarative agents. Sessions cover extending agents with skills and actions, grounding responses in enterprise context with WorkIQ, connecting MCP apps, and evaluating quality with Microsoft 365 Copilot Evals. The series includes presentations, demos, and live Q&A with Microsoft product and engineering teams. Attendees are advised to review declarative agents and Copilot extensibility concepts beforehand.

### Source excerpt

Building on Microsoft 365 Copilot? Here's your playbook. Declarative agents are quickly becoming one of the most exciting ways to extend Microsoft 365 Copilot and bring organizational knowledge, workflows, and tools directly into the flow of work. But as agent capabilities grow, so does the need for practical guidance: How do you build agents that [...] The post The Microsoft 365 Copilot Agent's Playbook: A Practical Livestream Series for Building Better Agents appeared first on Microsoft for Developers.

## How to test agent experience changes without shipping them

DevFeed: [How to test agent experience changes without shipping them](<https://devfeed.tech/articles/how-to-test-agent-experience-changes-without-shipping-them-23830.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/how-to-test-agent-experience-changes-without-shipping-them/>)

Author: Waldek Mastykarz, Garry Trinder

Published: 2026-07-21T07:15:12Z

Content type: article

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [product-experimentation](<https://devfeed.tech/topics/product-experimentation.md>), [agent observability](<https://devfeed.tech/topics/agent-observability.md>), [mock api](<https://devfeed.tech/topics/mock-api.md>), [Documentation](<https://devfeed.tech/topics/documentation.md>), [Model Context Protocol](<https://devfeed.tech/topics/model-context-protocol.md>), [MSP MCP](<https://devfeed.tech/topics/msp-mcp.md>), [Release notes](<https://devfeed.tech/topics/release-notes.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [ai](<https://devfeed.tech/tags/ai.md>), [api](<https://devfeed.tech/tags/api.md>), [ax](<https://devfeed.tech/tags/ax.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [can](<https://devfeed.tech/tags/can.md>), [changes](<https://devfeed.tech/tags/changes.md>), [documentation](<https://devfeed.tech/tags/documentation.md>), [experience](<https://devfeed.tech/tags/experience.md>), [first](<https://devfeed.tech/tags/first.md>), [learn](<https://devfeed.tech/tags/learn.md>), [mcp](<https://devfeed.tech/tags/mcp.md>), [mcp-server](<https://devfeed.tech/tags/mcp-server.md>), [microsoft](<https://devfeed.tech/tags/microsoft.md>), [production](<https://devfeed.tech/tags/production.md>), [project](<https://devfeed.tech/tags/project.md>), [server](<https://devfeed.tech/tags/server.md>), [shipping](<https://devfeed.tech/tags/shipping.md>), [test](<https://devfeed.tech/tags/test.md>), [upgrade](<https://devfeed.tech/tags/upgrade.md>), [works](<https://devfeed.tech/tags/works.md>)

### AI overview

In an SPFx upgrade scenario, documentation tips recommending a CLI were ignored across runs, while a warning that the agent's existing approach would cause build failures led it to use the CLI in all five runs. Other tests found that removing a competing migration guide reduced dependency accuracy from 34/50 to 18/50; structured JSON improved configuration correctness from 72/85 to 85/85, while JSONL cost 2.6 times more tokens and performed worse across dimensions; PowerShell examples harmed idiomatic code quality, and an explicit script was not executed in any of ten runs. The post recommends testing such hypotheses before shipping by intercepting agent requests and returning modified responses locally. It used Dev Proxy to emulate Microsoft Learn page changes and says the same approach can test MCP response formats. The described benefit is faster iteration--minutes rather than days--so teams can measure many hypotheses and ship changes that improve results.

### Source excerpt

Most changes you think will improve AI agent behavior won't. We tested a dozen hypotheses on a real project upgrade scenario and the majority failed. Learn how to emulate documentation, API, and MCP server changes locally so you can validate what works before shipping anything to production. The post How to test agent experience changes without shipping them appeared first on Microsoft for Developers.

## How to test agent skills without hitting real APIs

DevFeed: [How to test agent skills without hitting real APIs](<https://devfeed.tech/articles/how-to-test-agent-skills-without-hitting-real-apis-23831.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/how-to-test-agent-skills-without-hitting-real-apis/>)

Author: Waldek Mastykarz

Published: 2026-07-17T09:27:50Z

Content type: article

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [Agent Skills](<https://devfeed.tech/topics/agent-skills.md>), [mock api](<https://devfeed.tech/topics/mock-api.md>), [Software Testing](<https://devfeed.tech/topics/software-testing.md>), [Test automation](<https://devfeed.tech/topics/test-automation.md>), [Mocking](<https://devfeed.tech/topics/mocking.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [agent-skill](<https://devfeed.tech/tags/agent-skill.md>), [agent-skills](<https://devfeed.tech/tags/agent-skills.md>), [ai](<https://devfeed.tech/tags/ai.md>), [api](<https://devfeed.tech/tags/api.md>), [apis](<https://devfeed.tech/tags/apis.md>), [ax](<https://devfeed.tech/tags/ax.md>), [costs](<https://devfeed.tech/tags/costs.md>), [data](<https://devfeed.tech/tags/data.md>), [developers](<https://devfeed.tech/tags/developers.md>), [effect](<https://devfeed.tech/tags/effect.md>), [evals](<https://devfeed.tech/tags/evals.md>), [for-developers](<https://devfeed.tech/tags/for-developers.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [learn-how-to](<https://devfeed.tech/tags/learn-how-to.md>), [microsoft-for-developers](<https://devfeed.tech/tags/microsoft-for-developers.md>), [mock](<https://devfeed.tech/tags/mock.md>), [money](<https://devfeed.tech/tags/money.md>), [production](<https://devfeed.tech/tags/production.md>), [run](<https://devfeed.tech/tags/run.md>), [skills](<https://devfeed.tech/tags/skills.md>), [test](<https://devfeed.tech/tags/test.md>)

### AI overview

Evaluating API-calling agent skills against real services can incur call costs, mutate live data, and produce inconsistent results when outside changes affect shared state. The excerpt contrasts manually maintained mock servers, which need realistic endpoint behavior and can drift from the API, with Dev Proxy, configured through JSON to intercept the skill's unchanged real URLs and return seeded local data with relevant HTTP semantics. Its data resets between runs, avoiding production writes and API costs. The excerpt presents proxy emulation as a lighter option, while noting the choice depends on the situation; it does not cover the full setup.

### Source excerpt

Your agent skill calls an API. The moment you start evaluating it, every run either costs money or mutates production data. Learn how to mock APIs transparently so you can run evals without changing your skill or hitting real endpoints. The post How to test agent skills without hitting real APIs appeared first on Microsoft for Developers.

## Building AX evals that actually work

DevFeed: [Building AX evals that actually work](<https://devfeed.tech/articles/building-ax-evals-that-actually-work-23828.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/building-ax-evals-that-actually-work/>)

Author: Waldek Mastykarz

Published: 2026-07-15T12:53:13Z

Content type: article

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [Developer experience](<https://devfeed.tech/topics/developer-experience.md>), [cursor](<https://devfeed.tech/topics/cursor.md>), [Integration testing](<https://devfeed.tech/topics/integration-testing.md>), [Test-driven development](<https://devfeed.tech/topics/tdd.md>)

Tags: [adoption](<https://devfeed.tech/tags/adoption.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [agents](<https://devfeed.tech/tags/agents.md>), [aggregate](<https://devfeed.tech/tags/aggregate.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-coding-agents](<https://devfeed.tech/tags/ai-coding-agents.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [analyze](<https://devfeed.tech/tags/analyze.md>), [anecdotes](<https://devfeed.tech/tags/anecdotes.md>), [api](<https://devfeed.tech/tags/api.md>), [ax](<https://devfeed.tech/tags/ax.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [developers](<https://devfeed.tech/tags/developers.md>), [evals](<https://devfeed.tech/tags/evals.md>), [experience](<https://devfeed.tech/tags/experience.md>), [extensions](<https://devfeed.tech/tags/extensions.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [series](<https://devfeed.tech/tags/series.md>), [technology](<https://devfeed.tech/tags/technology.md>)

### AI overview

Reliable AX evals need realistic developer prompts, precise criteria that check correct usage and working code, and explicit pass, fail, and skip conditions. Keep prompts separate from rubrics, group empty-workspace and integration scenarios, and use clean environments representative of users' platforms. Run each scenario at least five times per condition; calibrate judges against known verdicts on real agent outputs and repeat calibration when criteria, models, presentation, or harnesses change. The article's examples show where to divide work: use a custom tool for semver comparisons and deterministic code for aggregate scores, while judges return individual verdicts with evidence. Start with 3-5 common tasks, pin and record the environment, version the rubric, and run on a cadence that leaves time to investigate changes. These recommendations and examples come from the supplied article text.

### Source excerpt

This is the eighth and final article in a series about Agent Experience (AX): the practice of making AI coding agents work correctly with your technology. The series covers what you can and can't control in the agent stack, how to measure whether your extensions are helping or hurting, and how to iterate toward better [...] The post Building AX evals that actually work appeared first on Microsoft for Developers.

## The hidden variables in your agent eval

DevFeed: [The hidden variables in your agent eval](<https://devfeed.tech/articles/the-hidden-variables-in-your-agent-eval-23834.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/blog/the-hidden-variables-in-your-agent-eval/>)

Author: Waldek Mastykarz

Published: 2026-07-08T12:11:17Z

Content type: article

Language: en

Sources: [Developer Blogs](<https://devfeed.tech/sources/developer-blogs.md>)

Topics: [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [Developer experience](<https://devfeed.tech/topics/developer-experience.md>), [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-experience](<https://devfeed.tech/tags/agent-experience.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-coding-agents](<https://devfeed.tech/tags/ai-coding-agents.md>), [api](<https://devfeed.tech/tags/api.md>), [auth](<https://devfeed.tech/tags/auth.md>), [authentication](<https://devfeed.tech/tags/authentication.md>), [ax](<https://devfeed.tech/tags/ax.md>), [azure](<https://devfeed.tech/tags/azure.md>), [bash](<https://devfeed.tech/tags/bash.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [bias](<https://devfeed.tech/tags/bias.md>), [build](<https://devfeed.tech/tags/build.md>), [eval](<https://devfeed.tech/tags/eval.md>), [extensions](<https://devfeed.tech/tags/extensions.md>), [hidden](<https://devfeed.tech/tags/hidden.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [net](<https://devfeed.tech/tags/net.md>), [stack](<https://devfeed.tech/tags/stack.md>), [variables](<https://devfeed.tech/tags/variables.md>)

### AI overview

Agent eval scores can shift because the model sees environmental context beyond the scenario: OS and shell, semantic clues in absolute paths, LSP diagnostics, dependencies, and changing agent or harness versions. The article reports that OS differences can affect shell fluency and even technology choices when prompts are ambiguous; path words such as "azure" or "poc" can bias provider choices or implementation rigor. LSP feedback can correct errors during generation and alter later turns, so stripped-down eval rigs may not reflect developer machines. To improve reproducibility, match the environment to the intended users, pin tools and LSP versions, use neutral workspace paths, choose deliberately whether dependencies are locked, and record an environment manifest. Some model-provider and harness updates cannot be fully pinned, so score changes need context.

### Source excerpt

This is the seventh article in a series about Agent Experience (AX): the practice of making AI coding agents work correctly with your technology. The series covers what you can and can't control in the agent stack, how to measure whether your extensions are helping or hurting, and how to iterate toward better outcomes. You [...] The post The hidden variables in your agent eval appeared first on Microsoft for Developers.

[Next page](<https://devfeed.tech/sources/developer-blogs.md?cursor=WyIyMDI2LTA3LTA4VDEyOjExOjE3KzAwOjAwIiwgImMxMzhjZTI3LTFmNzQtNDA3Yy1hM2MxLWZiMTk5YTI0OThlMSJd>)