feat(agents): opinionated defaults for azd ai agent init -m by v1212 · Pull Request #8426 · Azure/azure-dev

v1212 · 2026-05-28T07:52:11Z

Summary

Streamlined azd ai agent init experience when driven by a manifest (via -m <url> or interactive template selection). Introduces opinionated defaults that auto-resolve several decisions to reduce interactive prompts for quickstart scenarios.

Trigger: userProvidedManifest is true when the init flow is driven by a manifest -- either explicitly via -m flag/positional argument, or when the user interactively selects a template that resolves to a manifest.

About the -m URL: The manifest URL comes from azd ai agent sample list --output json (fields: manifestUrl, initCommand), dynamically sourced from the centrally-managed template catalog at https://aka.ms/foundry-agents-samples. Users simply copy it from the catalog or quickstart docs.

This PR is a prototype for team review -- validates the approach and UX before progressing to full product feature.

Before vs After

Before: `azd ai agent init -m <url>` (8+ prompts)

\
? How would you like to deploy your agent?

Container Image (Docker)
Source Code (ZIP upload)
? Select resources (CPU and Memory) for your agent
0.5 cores, 1Gi memory
? Select a Foundry project to host your agent
Use an existing Foundry project
? Select an Azure subscription
My Subscription (xxxxxxxx-...)
? Select a Foundry project
my-project (canadacentral)
? Model 'gpt-4.1-mini' is specified in the agent manifest.
Use 'gpt-4.1-mini' (from manifest)
? Found 1 existing deployment matching gpt-4.1-mini.
gpt-4.1-mini
? Select resources (CPU and Memory)
0.5 cores, 1Gi memory
\\

After: `azd ai agent init -m <url>` (3 prompts)

\
Downloading agent template...
? Select an Azure subscription

My Subscription (xxxxxxxx-...)
? Select a Foundry project
my-project (canadacentral)
? Model 'gpt-4.1-mini' -- select action:
Use existing deployment 'gpt-4.1-mini' (version: 2025-04-14)
Deploy a new model
Skip this model (do not deploy)

AI agent definition added to your azd project successfully!
\\

Prompt Count Comparison

Scenario	Before	After	Reduction
Existing project + 1 matching deployment	8	3	-5
Existing project + multiple deployments	8	3	-5
New project (no existing projects)	11-14	3	-8~11
init without -m (unchanged)	10-12	10-12	0

Auto-Resolution Logic

Step	Current Behavior	With manifest (this PR)	Override
Deploy mode	PROMPT: Container / Code	Auto: container	`--deploy-mode code`
Runtime	PROMPT: select	Auto-detect from project files	`--runtime <value>`
Entry point	PROMPT: enter path	Auto-detect (app.py / Program.cs)	`--entry-point <path>`
Container resources	PROMPT: select tier	Auto: smallest (0.5 cores, 1Gi)	N/A
Model confirm	PROMPT: Use from manifest?	Auto-confirm	N/A
Deployment match	PROMPT: select from list	PROMPT: Use/Change/Skip	`--no-prompt` auto-selects first
No-match deployment	PROMPT: Deploy new / Use different	Auto: deploy new	N/A

Key Design Decisions

Default deploy mode is container (not code) -- safer/more universal default; users who want code deploy pass --deploy-mode code
Model deployment always prompts Use/Change/Skip -- user should consciously choose which deployment to use (or skip/create new)
Applies to all manifest-driven flows -- both -m <url> and interactive template selection
Resolution precedence: explicit flag > !showCodeDeploy > userProvidedManifest > noPrompt > interactive
No silent magic -- auto-resolved values are logged (visible with --debug)

Implementation

userProvidedManifest field on InitAction -- captured when manifest is provided via -m or template selection
promptDeployMode: auto-selects container when manifest-driven
promptCodeConfig: auto-detects runtime + entry point
populateContainerSettings: auto-selects smallest tier
getModelDetails: skips model confirmation prompt
getModelDeploymentDetails: shows Use/Change/Skip prompt (no silent auto-select); auto-deploys-new when no match found
parseGitHubUrlNaive: refactored from method to package-level function (per review feedback)

Testing

All existing unit tests pass (13/13: TestPromptDeployMode_FlagOverride x5 + TestPromptCodeConfig_FlagOverrides x8)
6 new test cases for promptDeployMode and promptCodeConfig
356 lines of new tests for agent name resolution and manifest path handling (init_test.go)

E2E Test Results (2026-05-29, North Central US)

Tested on branch pr8426-latest with Foundry project wujia-aifproject-260527. All agents deployed and invoked successfully.

#	Scenario	Prompts Before	Prompts After	Reduction	Init	Deploy	Invoke	Result
1	Python `-m` + `--project-id`	8-12	0	-8~12	OK	OK (container, v6)	OK	PASS
2	C# `-m` + `--project-id`	8-12	0	-8~12	OK	OK (container, v2)	OK	PASS
3	`-m` without `--project-id` (interactive)	8-12	2-4	-6~8	OK	OK (container, v1)	OK	PASS
5	`-m` + `--project-id` + `--deploy-mode code --runtime --entry-point`	8-12	0	-8~12	OK	OK (code ZIP, v7)	OK	PASS
6	No `-m` flag (unchanged)	8-12	8-12	0	No regression	—	—	PASS

When users provide a manifest explicitly via -m flag or positional argument, apply opinionated defaults to minimize interactive prompts: - Deploy mode: auto-select 'code' for Python/.NET projects - Runtime: auto-detect from project files (requirements.txt python_3_13, .csproj dotnet_10) - Entry point: auto-detect using existing detectDefaultEntryPoint logic - Dependencies: default to remote_build - Model confirmation: skip 'Use from manifest / Choose different' prompt - Model deployment: auto-select if exactly one matching deployment exists - Container resources: auto-select default tier (0.5 cores, 1Gi) This reduces the -m quickstart from 8+ prompts to 2 (subscription + project). All auto-resolved values are printed with a prefix so users see what was chosen. The existing interactive init (without -m) and --no-prompt mode are completely unchanged. The userProvidedManifest bool is captured before auto-detection can set manifestPointer, ensuring only explicit user intent triggers the streamlined flow. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

- Print ' Model: <name> (from manifest)' when skipping model confirm prompt - Include version in deployment auto-select message for user clarity Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

- Auto-select 'deploy_new' when userProvidedManifest + no matching deployments (eliminates unnecessary prompt) - Document resolution precedence in promptDeployMode comment - Rename misleading test name to clarify what it actually tests Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…cope - Print checkmark when auto-selecting deploy_new for no-match scenario - Add comment explaining populateContainerSettings trigger conditions Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…n agent initialization

PR #8347 switched hosted-agent management calls from api-version=2025-11-15-preview to api-version=v1. The preview API version had implicitly satisfied the Foundry preview-feature gate, so requests went through without an opt-in header. The v1 endpoint enforces the gate explicitly: POST /agents/{name}/versions with definition.kind=="hosted" now returns HTTP 403 preview_feature_required unless the caller sends 'Foundry-Features: HostedAgents=V1Preview'. Effect on users: 'azd deploy' fails for container (hosted) agents with the preview_feature_required error returned from the Foundry API. Confirmed via direct API probes against a live Foundry project: same URL/body/auth, only the api-version and header differ. api-version=2025-11-15-preview, no header -> 200 OK api-version=v1, no header -> 403 preview_feature_required api-version=v1, with header -> 200 OK Add the header inside CreateAgentVersion so all callers get the fix (service_target_agent.go deployHostedAgent and cmd/optimize_deploy.go). Scope: only CreateAgentVersion needs the change. Probed sibling v1 management calls (GetAgent, GetAgentVersion, ListAgents, ListAgentVersions, PatchAgent) against the same live project without the header - none returned 403. The write path that creates a hosted resource is the only one currently gated. Add a focused test that asserts CreateAgentVersion always sets the header, so this can't silently regress again. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…ilently auto-selecting When `azd ai agent init -m <manifest>` selected a Foundry project that already contained exactly one matching deployment for the manifest's model resource, the code silently auto-selected that deployment with no prompt. That made the model an invisible default and was inconsistent with the Use/Change/Skip selector shown everywhere else. This change replaces the silent path and the older "Found N existing deployment(s)... Create new model deployment" selector with a unified Use/Change/Skip-style prompt: - "Use existing deployment 'name' (version: X)" (one per matching deployment, sorted by name for deterministic ordering) - "Deploy a new model" (falls through to the existing deploy-new path which loads the model catalog) - "Skip this model (do not deploy)" (returns errModelSkipped; the resource is dropped from manifest.Resources by ProcessModels) The selector uses stable `SelectChoice.Value` strings ("use:<name>", "deploy_new", "skip") so behavior is keyed off values, not labels. In `--no-prompt` mode the first sorted matching deployment is selected automatically (no prompt) so headless/CI flows continue to work. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Signed-off-by: trangevi <trangevi@microsoft.com>

Copilot

Pull request overview

This PR prototypes a streamlined azd ai agent init -m <manifest> experience by introducing “opinionated defaults” that auto-resolve several decisions (deploy mode/config, model selection, container resources) to reduce interactive prompts for quickstart scenarios.

Changes:

Add userProvidedManifest plumbing through the init flow, plus manifest “peeking” helpers to derive agent/project folder naming earlier.
Adjust deploy-mode / code-config prompting behavior to support auto-resolution paths.
Refactor model deployment selection to be deterministic in --no-prompt, add “skip model” support, and update tests accordingly.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 10 comments.

Show a summary per file

File	Description
`cli/azd/extensions/azure.ai.agents/internal/cmd/init.go`	Adds manifest-peek + early agent-name resolution, threads `userProvidedManifest`, changes container tier prompting, and alters some stdout/log output behavior.
`cli/azd/extensions/azure.ai.agents/internal/cmd/init_test.go`	Adds unit tests covering manifest peek + early agent-name resolution + manifest-path absolutization.
`cli/azd/extensions/azure.ai.agents/internal/cmd/init_models.go`	Adds model-skip sentinel, refactors deployment selection UI, and adjusts model processing to drop skipped resources.
`cli/azd/extensions/azure.ai.agents/internal/cmd/init_from_code.go`	Extends deploy-mode and code-config prompting to accept `userProvidedManifest` and apply defaults.
`cli/azd/extensions/azure.ai.agents/internal/cmd/init_from_code_test.go`	Adds coverage for the new `userProvidedManifest` behavior in deploy-mode/code-config prompting.

…l function directly Per review feedback (trangevi): the InitAction.parseGitHubUrlNaive() method was just a passthrough to the package-level function. Remove the wrapper and call the package-level function directly. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

github-actions · 2026-05-29T02:53:01Z

📋 Prioritization Note

Thanks for the contribution! The linked issue isn't in the current milestone yet.
Review may take a bit longer — reach out to @rajeshkamal5050 or @kristenwomack if you'd like to discuss prioritization.

trangevi · 2026-05-29T14:33:42Z

/check-enforcer override

microsoft-github-policy-service Bot assigned v1212 May 28, 2026

github-actions Bot added the ext-agents azure.ai.{agents,connections,inspector,projects,routines,skills,toolboxes} extensions label May 28, 2026

Jian Wu and others added 11 commits May 28, 2026 16:00

fix: add output for model confirm and version in deployment auto-select

9412788

- Print ' Model: <name> (from manifest)' when skipping model confirm prompt - Include version in deployment auto-select message for user clarity Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

refactor: replace fmt.Printf with log.Printf for consistent logging i…

b5cf2d1

…n agent initialization

feat(models): allow skipping model selection during deployment process

96e8a84

refactor: default to container deploy not code deploy

e72ff42

Merge branch 'main' into jw/init-opinionated-defaults

06a74b2

Use opinionated defaults for all manifest paths

abf968c

Signed-off-by: trangevi <trangevi@microsoft.com>

feat(init): implement agent name resolution and manifest path handling

eb8b0b3

v1212 marked this pull request as ready for review May 29, 2026 02:06

Copilot AI review requested due to automatic review settings May 29, 2026 02:06

v1212 requested review from JeffreyCA, therealjohn, trangevi and trrwilson as code owners May 29, 2026 02:06

Copilot started reviewing on behalf of v1212 May 29, 2026 02:06 View session

Copilot AI reviewed May 29, 2026

View reviewed changes

trangevi requested changes May 29, 2026

View reviewed changes

Comment thread cli/azd/extensions/azure.ai.agents/internal/cmd/init.go Outdated

trangevi approved these changes May 29, 2026

View reviewed changes

trangevi enabled auto-merge (squash) May 29, 2026 14:33

trangevi merged commit 5d89dd7 into main May 29, 2026
28 of 31 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat(agents): opinionated defaults for azd ai agent init -m#8426

feat(agents): opinionated defaults for azd ai agent init -m#8426
trangevi merged 13 commits into
mainfrom
jw/init-opinionated-defaults

v1212 commented May 28, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions Bot commented May 29, 2026 •

edited

Loading

Uh oh!

trangevi commented May 29, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

v1212 commented May 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Before vs After

Before: azd ai agent init -m <url> (8+ prompts)

After: azd ai agent init -m <url> (3 prompts)

Prompt Count Comparison

Auto-Resolution Logic

Key Design Decisions

Implementation

Testing

E2E Test Results (2026-05-29, North Central US)

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions Bot commented May 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

📋 Prioritization Note

Uh oh!

trangevi commented May 29, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

v1212 commented May 28, 2026 •

edited

Loading

Before: `azd ai agent init -m <url>` (8+ prompts)

After: `azd ai agent init -m <url>` (3 prompts)

github-actions Bot commented May 29, 2026 •

edited

Loading