fix: distinguish GitHub Copilot App execution environment

[qinezh]

Summary

• detect the GitHub Copilot App from the exact AI_AGENT=github_copilot_app_agent marker before Copilot CLI fallback markers
• report GitHub Copilot App through execution.environment
• add precedence coverage and isolate tests from the ambient AI_AGENT variable

Testing

• go test ./internal/runcontext/agentdetect ./internal/tracing/resource ./internal/terminal
• go test ./cmd -run '^(TestParseGlobalFlags_AgentDetection|TestAgentDetectionIntegration)$'

Fixes #9289

[azure-pipelines]

Azure Pipelines:
7 pipeline(s) were filtered out due to trigger conditions.
There may be pipelines that require an authorized user to comment /azp run to run.

[Copilot]

Pull request overview

Adds distinct GitHub Copilot App detection and telemetry reporting.

Changes:

• Detects the exact AI_AGENT marker with CLI-marker precedence.
• Adds agent and telemetry environment constants.
• Expands detection coverage and test isolation.

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
docs/reference/telemetry-data.md	Documents the telemetry value.
cli/azd/internal/tracing/resource/exec_environment.go	Maps the detected app to telemetry.
cli/azd/internal/tracing/fields/fields.go	Defines the telemetry value.
cli/azd/internal/terminal/terminal_test.go	Clears ambient AI_AGENT.
cli/azd/internal/runcontext/agentdetect/types.go	Defines the app agent type.
cli/azd/internal/runcontext/agentdetect/detect_test.go	Tests detection and precedence.
cli/azd/internal/runcontext/agentdetect/detect_env.go	Detects the exact app marker.
cli/azd/cmd/auto_install_integration_test.go	Isolates integration tests from AI_AGENT.

[jongio]

Reviewed the change. The exact AI_AGENT=github_copilot_app_agent match is gated ahead of the Copilot CLI markers and wired through both DisplayName and execEnvFromAgent. Precedence and the unrecognized-value case are covered by the new tests, and the targeted suites pass locally (agentdetect, tracing/resource, terminal, and the cmd agent-detection tests).

The three open automated-review comments look non-blocking:

• environment-variables.md: existing host-detection markers like COPILOT_CLI and CLAUDE_CODE aren't listed there either, so omitting AI_AGENT stays consistent with the current convention.
• telemetry-schema.md: it documents fields, not the execution.environment value list. That enumeration lives in telemetry-data.md, which this PR updates.
• resource-level test: there's no per-agent execEnvFromAgent test today, and the behavior being fixed is the detection precedence, which is already covered.

[JeffreyCA]

Thank you, looks good!

[JeffreyCA]

/check-enforcer evaluate