fix(deploy): pass stack selection to diff for --target filtering

[Hweinstock]

Description

When --diff or --deploy is used with --target, the CLI now passes a stack selection pattern (*${targetName}*) to the CDK toolkit so only the selected target's stack is operated on, not all stacks in the synthesized app.

Extracted runDiff and runDeploy functions for testability via dependency injection.

Related Issue

Closes#980

Type of Change

• Bug fix

Testing

• I ran npm run test:unit and npm run test:integ
• I ran npm run typecheck
• I ran npm run lint

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• My changes generate no new warnings

[github-actions]

Coverage Report

Status	Category	Percentage	Covered / Total
🔵	Lines	43.19%	9035 / 20915
🔵	Statements	42.47%	9594 / 22585
🔵	Functions	40.07%	1560 / 3893
🔵	Branches	40%	5811 / 14527

Generated in workflow #2576 for commit f95c643 by the Vitest Coverage Report Action

[agentcore-cli-automation]

The fix passes stackName into runDiff, but stackName is defined at line 239 as:

conststackName=stackNames[0]!;

stackNames contains all stacks synthesized by the CDK app, and bin/cdk.ts creates one stack per entry in aws-targets.json (AgentCore-${project}-${target}). stackNames[0] is therefore just the first stack in the list (effectively arbitrary ordering), not the stack that corresponds to options.target.

Reproduction of the original bug #980:

• Targets in aws-targets.json: preprod-east2, default
• Run agentcore deploy --target default --diff
• stackNames = ['AgentCore-MyProject-preprod-east2', 'AgentCore-MyProject-default'] (or whatever order CDK returns)
• stackNames[0] = 'AgentCore-MyProject-preprod-east2'
• Diff runs against the wrong stack — the exact issue this PR is trying to fix.

Even in the current test setup this passes because the test hardcodes 'MyProject-prod' as the stack name and only asserts that the wrapper receives whatever string was passed in. It doesn't verify that handleDeploy selects the correct stack for options.target.

Suggested fixes (pick one):

1. Construct the expected stack name from spec.name + target.name using the same sanitize logic as src/assets/cdk/bin/cdk.ts (AgentCore-${sanitize(projectName)}-${sanitize(targetName)}) and find it in stackNames, erroring if it's missing.
2. Filter stackNames to the one that ends with -${sanitize(target.name)} (less robust, but simple).
3. Pass target.name (or both project + target) into runDiff and let it build the pattern itself.

It's probably also worth adding a test that exercises handleDeploy with multiple synthesized stacks and asserts the diff is scoped to the stack matching options.target — otherwise a future regression here won't be caught.

As a side note, the issue description also flags that toolkitWrapper.deploy() at line 334 has the same missing stack selection. That's out of scope for this PR, but worth mentioning since both are driven by the same underlying stackNames[0] assumption.

[Hweinstock]

Addressed by no longer relying on arbitrary first stack name behavior.

[Hweinstock]

Also fixes #1150

[jesseturner21]

good fix!

[Hweinstock]

Description

When --diff or --deploy is used with --target, the CLI now passes a stack selection pattern (*${targetName}*) to the CDK toolkit so only the selected target's stack is operated on, not all stacks in the synthesized app.

Extracted runDiff and runDeploy functions for testability via dependency injection.

Related Issue

Closes#980

Type of Change

• Bug fix

Testing

• I ran npm run test:unit and npm run test:integ
• I ran npm run typecheck
• I ran npm run lint

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• My changes generate no new warnings

[github-actions]

Coverage Report

Status	Category	Percentage	Covered / Total
🔵	Lines	43.19%	9035 / 20915
🔵	Statements	42.47%	9594 / 22585
🔵	Functions	40.07%	1560 / 3893
🔵	Branches	40%	5811 / 14527

Generated in workflow #2576 for commit f95c643 by the Vitest Coverage Report Action

[agentcore-cli-automation]

The fix passes stackName into runDiff, but stackName is defined at line 239 as:

conststackName=stackNames[0]!;

stackNames contains all stacks synthesized by the CDK app, and bin/cdk.ts creates one stack per entry in aws-targets.json (AgentCore-${project}-${target}). stackNames[0] is therefore just the first stack in the list (effectively arbitrary ordering), not the stack that corresponds to options.target.

Reproduction of the original bug #980:

• Targets in aws-targets.json: preprod-east2, default
• Run agentcore deploy --target default --diff
• stackNames = ['AgentCore-MyProject-preprod-east2', 'AgentCore-MyProject-default'] (or whatever order CDK returns)
• stackNames[0] = 'AgentCore-MyProject-preprod-east2'
• Diff runs against the wrong stack — the exact issue this PR is trying to fix.

Even in the current test setup this passes because the test hardcodes 'MyProject-prod' as the stack name and only asserts that the wrapper receives whatever string was passed in. It doesn't verify that handleDeploy selects the correct stack for options.target.

Suggested fixes (pick one):

1. Construct the expected stack name from spec.name + target.name using the same sanitize logic as src/assets/cdk/bin/cdk.ts (AgentCore-${sanitize(projectName)}-${sanitize(targetName)}) and find it in stackNames, erroring if it's missing.
2. Filter stackNames to the one that ends with -${sanitize(target.name)} (less robust, but simple).
3. Pass target.name (or both project + target) into runDiff and let it build the pattern itself.

It's probably also worth adding a test that exercises handleDeploy with multiple synthesized stacks and asserts the diff is scoped to the stack matching options.target — otherwise a future regression here won't be caught.

As a side note, the issue description also flags that toolkitWrapper.deploy() at line 334 has the same missing stack selection. That's out of scope for this PR, but worth mentioning since both are driven by the same underlying stackNames[0] assumption.

[Hweinstock]

Addressed by no longer relying on arbitrary first stack name behavior.

[Hweinstock]

Also fixes #1150

[jesseturner21]

good fix!

[Hweinstock]

Description

When --diff or --deploy is used with --target, the CLI now passes a stack selection pattern (*${targetName}*) to the CDK toolkit so only the selected target's stack is operated on, not all stacks in the synthesized app.

Extracted runDiff and runDeploy functions for testability via dependency injection.

Related Issue

Closes#980

Type of Change

• Bug fix

Testing

• I ran npm run test:unit and npm run test:integ
• I ran npm run typecheck
• I ran npm run lint

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• My changes generate no new warnings

[github-actions]

Coverage Report

Status	Category	Percentage	Covered / Total
🔵	Lines	43.19%	9035 / 20915
🔵	Statements	42.47%	9594 / 22585
🔵	Functions	40.07%	1560 / 3893
🔵	Branches	40%	5811 / 14527

Generated in workflow #2576 for commit f95c643 by the Vitest Coverage Report Action

[agentcore-cli-automation]

The fix passes stackName into runDiff, but stackName is defined at line 239 as:

conststackName=stackNames[0]!;

stackNames contains all stacks synthesized by the CDK app, and bin/cdk.ts creates one stack per entry in aws-targets.json (AgentCore-${project}-${target}). stackNames[0] is therefore just the first stack in the list (effectively arbitrary ordering), not the stack that corresponds to options.target.

Reproduction of the original bug #980:

• Targets in aws-targets.json: preprod-east2, default
• Run agentcore deploy --target default --diff
• stackNames = ['AgentCore-MyProject-preprod-east2', 'AgentCore-MyProject-default'] (or whatever order CDK returns)
• stackNames[0] = 'AgentCore-MyProject-preprod-east2'
• Diff runs against the wrong stack — the exact issue this PR is trying to fix.

Even in the current test setup this passes because the test hardcodes 'MyProject-prod' as the stack name and only asserts that the wrapper receives whatever string was passed in. It doesn't verify that handleDeploy selects the correct stack for options.target.

Suggested fixes (pick one):

1. Construct the expected stack name from spec.name + target.name using the same sanitize logic as src/assets/cdk/bin/cdk.ts (AgentCore-${sanitize(projectName)}-${sanitize(targetName)}) and find it in stackNames, erroring if it's missing.
2. Filter stackNames to the one that ends with -${sanitize(target.name)} (less robust, but simple).
3. Pass target.name (or both project + target) into runDiff and let it build the pattern itself.

It's probably also worth adding a test that exercises handleDeploy with multiple synthesized stacks and asserts the diff is scoped to the stack matching options.target — otherwise a future regression here won't be caught.

As a side note, the issue description also flags that toolkitWrapper.deploy() at line 334 has the same missing stack selection. That's out of scope for this PR, but worth mentioning since both are driven by the same underlying stackNames[0] assumption.

[Hweinstock]

Addressed by no longer relying on arbitrary first stack name behavior.

[Hweinstock]

Also fixes #1150

[jesseturner21]

good fix!

[Hweinstock]

Description

When --diff or --deploy is used with --target, the CLI now passes a stack selection pattern (*${targetName}*) to the CDK toolkit so only the selected target's stack is operated on, not all stacks in the synthesized app.

Extracted runDiff and runDeploy functions for testability via dependency injection.

Related Issue

Closes#980

Type of Change

• Bug fix

Testing

• I ran npm run test:unit and npm run test:integ
• I ran npm run typecheck
• I ran npm run lint

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• My changes generate no new warnings

[github-actions]

Coverage Report

Status	Category	Percentage	Covered / Total
🔵	Lines	43.19%	9035 / 20915
🔵	Statements	42.47%	9594 / 22585
🔵	Functions	40.07%	1560 / 3893
🔵	Branches	40%	5811 / 14527

Generated in workflow #2576 for commit f95c643 by the Vitest Coverage Report Action

[agentcore-cli-automation]

The fix passes stackName into runDiff, but stackName is defined at line 239 as:

conststackName=stackNames[0]!;

stackNames contains all stacks synthesized by the CDK app, and bin/cdk.ts creates one stack per entry in aws-targets.json (AgentCore-${project}-${target}). stackNames[0] is therefore just the first stack in the list (effectively arbitrary ordering), not the stack that corresponds to options.target.

Reproduction of the original bug #980:

• Targets in aws-targets.json: preprod-east2, default
• Run agentcore deploy --target default --diff
• stackNames = ['AgentCore-MyProject-preprod-east2', 'AgentCore-MyProject-default'] (or whatever order CDK returns)
• stackNames[0] = 'AgentCore-MyProject-preprod-east2'
• Diff runs against the wrong stack — the exact issue this PR is trying to fix.

Even in the current test setup this passes because the test hardcodes 'MyProject-prod' as the stack name and only asserts that the wrapper receives whatever string was passed in. It doesn't verify that handleDeploy selects the correct stack for options.target.

Suggested fixes (pick one):

1. Construct the expected stack name from spec.name + target.name using the same sanitize logic as src/assets/cdk/bin/cdk.ts (AgentCore-${sanitize(projectName)}-${sanitize(targetName)}) and find it in stackNames, erroring if it's missing.
2. Filter stackNames to the one that ends with -${sanitize(target.name)} (less robust, but simple).
3. Pass target.name (or both project + target) into runDiff and let it build the pattern itself.

It's probably also worth adding a test that exercises handleDeploy with multiple synthesized stacks and asserts the diff is scoped to the stack matching options.target — otherwise a future regression here won't be caught.

As a side note, the issue description also flags that toolkitWrapper.deploy() at line 334 has the same missing stack selection. That's out of scope for this PR, but worth mentioning since both are driven by the same underlying stackNames[0] assumption.

[Hweinstock]

Addressed by no longer relying on arbitrary first stack name behavior.

[Hweinstock]

Also fixes #1150

[jesseturner21]

good fix!

[Hweinstock]

Description

When --diff or --deploy is used with --target, the CLI now passes a stack selection pattern (*${targetName}*) to the CDK toolkit so only the selected target's stack is operated on, not all stacks in the synthesized app.

Extracted runDiff and runDeploy functions for testability via dependency injection.

Related Issue

Closes#980

Type of Change

• Bug fix

Testing

• I ran npm run test:unit and npm run test:integ
• I ran npm run typecheck
• I ran npm run lint

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• My changes generate no new warnings

[github-actions]

Coverage Report

Status	Category	Percentage	Covered / Total
🔵	Lines	43.19%	9035 / 20915
🔵	Statements	42.47%	9594 / 22585
🔵	Functions	40.07%	1560 / 3893
🔵	Branches	40%	5811 / 14527

Generated in workflow #2576 for commit f95c643 by the Vitest Coverage Report Action

[agentcore-cli-automation]

The fix passes stackName into runDiff, but stackName is defined at line 239 as:

conststackName=stackNames[0]!;

stackNames contains all stacks synthesized by the CDK app, and bin/cdk.ts creates one stack per entry in aws-targets.json (AgentCore-${project}-${target}). stackNames[0] is therefore just the first stack in the list (effectively arbitrary ordering), not the stack that corresponds to options.target.

Reproduction of the original bug #980:

• Targets in aws-targets.json: preprod-east2, default
• Run agentcore deploy --target default --diff
• stackNames = ['AgentCore-MyProject-preprod-east2', 'AgentCore-MyProject-default'] (or whatever order CDK returns)
• stackNames[0] = 'AgentCore-MyProject-preprod-east2'
• Diff runs against the wrong stack — the exact issue this PR is trying to fix.

Even in the current test setup this passes because the test hardcodes 'MyProject-prod' as the stack name and only asserts that the wrapper receives whatever string was passed in. It doesn't verify that handleDeploy selects the correct stack for options.target.

Suggested fixes (pick one):

1. Construct the expected stack name from spec.name + target.name using the same sanitize logic as src/assets/cdk/bin/cdk.ts (AgentCore-${sanitize(projectName)}-${sanitize(targetName)}) and find it in stackNames, erroring if it's missing.
2. Filter stackNames to the one that ends with -${sanitize(target.name)} (less robust, but simple).
3. Pass target.name (or both project + target) into runDiff and let it build the pattern itself.

It's probably also worth adding a test that exercises handleDeploy with multiple synthesized stacks and asserts the diff is scoped to the stack matching options.target — otherwise a future regression here won't be caught.

As a side note, the issue description also flags that toolkitWrapper.deploy() at line 334 has the same missing stack selection. That's out of scope for this PR, but worth mentioning since both are driven by the same underlying stackNames[0] assumption.

[Hweinstock]

Addressed by no longer relying on arbitrary first stack name behavior.

[Hweinstock]

Also fixes #1150

[jesseturner21]

good fix!

[Hweinstock]

Description

When --diff or --deploy is used with --target, the CLI now passes a stack selection pattern (*${targetName}*) to the CDK toolkit so only the selected target's stack is operated on, not all stacks in the synthesized app.

Extracted runDiff and runDeploy functions for testability via dependency injection.

Related Issue

Closes#980

Type of Change

• Bug fix

Testing

• I ran npm run test:unit and npm run test:integ
• I ran npm run typecheck
• I ran npm run lint

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• My changes generate no new warnings

[github-actions]

Coverage Report

Status	Category	Percentage	Covered / Total
🔵	Lines	43.19%	9035 / 20915
🔵	Statements	42.47%	9594 / 22585
🔵	Functions	40.07%	1560 / 3893
🔵	Branches	40%	5811 / 14527

Generated in workflow #2576 for commit f95c643 by the Vitest Coverage Report Action

[agentcore-cli-automation]

The fix passes stackName into runDiff, but stackName is defined at line 239 as:

conststackName=stackNames[0]!;

stackNames contains all stacks synthesized by the CDK app, and bin/cdk.ts creates one stack per entry in aws-targets.json (AgentCore-${project}-${target}). stackNames[0] is therefore just the first stack in the list (effectively arbitrary ordering), not the stack that corresponds to options.target.

Reproduction of the original bug #980:

• Targets in aws-targets.json: preprod-east2, default
• Run agentcore deploy --target default --diff
• stackNames = ['AgentCore-MyProject-preprod-east2', 'AgentCore-MyProject-default'] (or whatever order CDK returns)
• stackNames[0] = 'AgentCore-MyProject-preprod-east2'
• Diff runs against the wrong stack — the exact issue this PR is trying to fix.

Even in the current test setup this passes because the test hardcodes 'MyProject-prod' as the stack name and only asserts that the wrapper receives whatever string was passed in. It doesn't verify that handleDeploy selects the correct stack for options.target.

Suggested fixes (pick one):

1. Construct the expected stack name from spec.name + target.name using the same sanitize logic as src/assets/cdk/bin/cdk.ts (AgentCore-${sanitize(projectName)}-${sanitize(targetName)}) and find it in stackNames, erroring if it's missing.
2. Filter stackNames to the one that ends with -${sanitize(target.name)} (less robust, but simple).
3. Pass target.name (or both project + target) into runDiff and let it build the pattern itself.

It's probably also worth adding a test that exercises handleDeploy with multiple synthesized stacks and asserts the diff is scoped to the stack matching options.target — otherwise a future regression here won't be caught.

As a side note, the issue description also flags that toolkitWrapper.deploy() at line 334 has the same missing stack selection. That's out of scope for this PR, but worth mentioning since both are driven by the same underlying stackNames[0] assumption.

[Hweinstock]

Addressed by no longer relying on arbitrary first stack name behavior.

[Hweinstock]

Also fixes #1150

[jesseturner21]

good fix!

[Hweinstock]

Description

When --diff or --deploy is used with --target, the CLI now passes a stack selection pattern (*${targetName}*) to the CDK toolkit so only the selected target's stack is operated on, not all stacks in the synthesized app.

Extracted runDiff and runDeploy functions for testability via dependency injection.

Related Issue

Closes#980

Type of Change

• Bug fix

Testing

• I ran npm run test:unit and npm run test:integ
• I ran npm run typecheck
• I ran npm run lint

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• My changes generate no new warnings

[github-actions]

Coverage Report

Status	Category	Percentage	Covered / Total
🔵	Lines	43.19%	9035 / 20915
🔵	Statements	42.47%	9594 / 22585
🔵	Functions	40.07%	1560 / 3893
🔵	Branches	40%	5811 / 14527

Generated in workflow #2576 for commit f95c643 by the Vitest Coverage Report Action

[agentcore-cli-automation]

The fix passes stackName into runDiff, but stackName is defined at line 239 as:

conststackName=stackNames[0]!;

stackNames contains all stacks synthesized by the CDK app, and bin/cdk.ts creates one stack per entry in aws-targets.json (AgentCore-${project}-${target}). stackNames[0] is therefore just the first stack in the list (effectively arbitrary ordering), not the stack that corresponds to options.target.

Reproduction of the original bug #980:

• Targets in aws-targets.json: preprod-east2, default
• Run agentcore deploy --target default --diff
• stackNames = ['AgentCore-MyProject-preprod-east2', 'AgentCore-MyProject-default'] (or whatever order CDK returns)
• stackNames[0] = 'AgentCore-MyProject-preprod-east2'
• Diff runs against the wrong stack — the exact issue this PR is trying to fix.

Even in the current test setup this passes because the test hardcodes 'MyProject-prod' as the stack name and only asserts that the wrapper receives whatever string was passed in. It doesn't verify that handleDeploy selects the correct stack for options.target.

Suggested fixes (pick one):

1. Construct the expected stack name from spec.name + target.name using the same sanitize logic as src/assets/cdk/bin/cdk.ts (AgentCore-${sanitize(projectName)}-${sanitize(targetName)}) and find it in stackNames, erroring if it's missing.
2. Filter stackNames to the one that ends with -${sanitize(target.name)} (less robust, but simple).
3. Pass target.name (or both project + target) into runDiff and let it build the pattern itself.

It's probably also worth adding a test that exercises handleDeploy with multiple synthesized stacks and asserts the diff is scoped to the stack matching options.target — otherwise a future regression here won't be caught.

As a side note, the issue description also flags that toolkitWrapper.deploy() at line 334 has the same missing stack selection. That's out of scope for this PR, but worth mentioning since both are driven by the same underlying stackNames[0] assumption.

[Hweinstock]

Addressed by no longer relying on arbitrary first stack name behavior.

[Hweinstock]

Also fixes #1150

[jesseturner21]

good fix!

[Hweinstock]

Description

When --diff or --deploy is used with --target, the CLI now passes a stack selection pattern (*${targetName}*) to the CDK toolkit so only the selected target's stack is operated on, not all stacks in the synthesized app.

Extracted runDiff and runDeploy functions for testability via dependency injection.

Related Issue

Closes#980

Type of Change

• Bug fix

Testing

• I ran npm run test:unit and npm run test:integ
• I ran npm run typecheck
• I ran npm run lint

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• My changes generate no new warnings

[github-actions]

Coverage Report

Status	Category	Percentage	Covered / Total
🔵	Lines	43.19%	9035 / 20915
🔵	Statements	42.47%	9594 / 22585
🔵	Functions	40.07%	1560 / 3893
🔵	Branches	40%	5811 / 14527

Generated in workflow #2576 for commit f95c643 by the Vitest Coverage Report Action

[agentcore-cli-automation]

The fix passes stackName into runDiff, but stackName is defined at line 239 as:

conststackName=stackNames[0]!;

stackNames contains all stacks synthesized by the CDK app, and bin/cdk.ts creates one stack per entry in aws-targets.json (AgentCore-${project}-${target}). stackNames[0] is therefore just the first stack in the list (effectively arbitrary ordering), not the stack that corresponds to options.target.

Reproduction of the original bug #980:

• Targets in aws-targets.json: preprod-east2, default
• Run agentcore deploy --target default --diff
• stackNames = ['AgentCore-MyProject-preprod-east2', 'AgentCore-MyProject-default'] (or whatever order CDK returns)
• stackNames[0] = 'AgentCore-MyProject-preprod-east2'
• Diff runs against the wrong stack — the exact issue this PR is trying to fix.

Even in the current test setup this passes because the test hardcodes 'MyProject-prod' as the stack name and only asserts that the wrapper receives whatever string was passed in. It doesn't verify that handleDeploy selects the correct stack for options.target.

Suggested fixes (pick one):

1. Construct the expected stack name from spec.name + target.name using the same sanitize logic as src/assets/cdk/bin/cdk.ts (AgentCore-${sanitize(projectName)}-${sanitize(targetName)}) and find it in stackNames, erroring if it's missing.
2. Filter stackNames to the one that ends with -${sanitize(target.name)} (less robust, but simple).
3. Pass target.name (or both project + target) into runDiff and let it build the pattern itself.

It's probably also worth adding a test that exercises handleDeploy with multiple synthesized stacks and asserts the diff is scoped to the stack matching options.target — otherwise a future regression here won't be caught.

As a side note, the issue description also flags that toolkitWrapper.deploy() at line 334 has the same missing stack selection. That's out of scope for this PR, but worth mentioning since both are driven by the same underlying stackNames[0] assumption.

[Hweinstock]

Addressed by no longer relying on arbitrary first stack name behavior.

[Hweinstock]

Also fixes #1150

[jesseturner21]

good fix!