NEW @W-21910054@ Message sanitization

[nikhil-mittal-165]

Summary

Sanitizes the run command's org-authentication error path so that ApexGuru auth failures are handled consistently and gracefully instead of aborting the whole run with a generic SfError.

• Removed RunAction.validateTargetOrg() / the upfront Org.create() pre-check. ApexGuru is the only engine that consumes --target-org, and ApexGuruAuthService/ApexGuruEngine (see code-analyzer-core#500) already produce a clear, actionable message and skip gracefully with a warning (NO_ORG_CONNECTION/INVALID_SESSION) rather than throwing. Keeping the CLI-level pre-check meant two divergent error messages for the same failure.
• Behavior/exit-code change (intentional): as of this PR, running sf code-analyzer run --target-org <bad-or-unauthenticated-org> no longer fails the whole command. ApexGuru is skipped with a warning (visible in the terminal output/log) and the command otherwise completes normally (exit code reflects violations found by other engines, not the org-auth failure). This is a deliberate consequence of delegating auth handling to the engine layer, but it does change exit-code/CI semantics for existing scripts that relied on --target-org failing fast on a bad org — calling it out explicitly here since it isn't just a message-wording change.
• Also bundles in ActionSummaryViewer.displayApexGuruAnalysisMode() (new run-action.apexguru-analysis-mode message), which prints ApexGuru analysis mode: ADVANCED/BASIC in the terminal summary when ApexGuru insights report an analysis mode. This is unrelated new functionality rather than sanitization — including it here rather than a separate PR since it's small, well-tested, and low-risk; flagging it explicitly per team convention so reviewers aren't surprised by scope.

Test plan

• test/lib/actions/RunAction.test.ts updated: removed the two tests asserting the old throw-on-bad-org behavior (with a comment explaining why, and pointing to core's ApexGuruAuthService/ApexGuruEngine tests for that coverage), added coverage for the new ApexGuru analysis-mode summary line (full→ADVANCED, static→BASIC, no line when absent/engine didn't run).
• Unit tests, lint, and typecheck pass locally.

[aruntyagiTutu]

The sanitization-motivated change is sound: removing RunAction.validateTargetOrg()/Org.create() pre-check makes sense once ApexGuruAuthService (code-analyzer-core#500) produces its own clear, actionable message and the engine skips gracefully with a warning instead of the CLI aborting the whole run with a generic SfError. That avoids two divergent error messages for the same failure and is well covered by the updated RunAction.test.ts (including the new comment explaining why the org-auth tests were removed).

One thing worth calling out since the PR description is empty (same pattern as sfdx-code-analyzer-vscode#368 and code-analyzer-core#500 from the same author): this PR also bundles in an unrelated feature — ActionSummaryViewer.displayApexGuruAnalysisMode() and the new run-action.apexguru-analysis-mode message, which prints ApexGuru analysis mode: ADVANCED/BASIC in the terminal summary. That's new user-facing functionality, not message sanitization, and per the team's "one concern per PR" convention it'd be cleaner as a separate PR. Not blocking since it's well-tested and low-risk, but worth a mention in the description (or splitting out) so reviewers aren't surprised by scope.

Also worth confirming: is it intentional that a bad/unauthenticated --target-org now causes the whole run command to succeed (ApexGuru skipped with a warning) rather than fail? That's a behavior change beyond message wording — likely the right call given ApexGuru is the only consumer of target-org, but since it changes exit-code/CI semantics for existing users of --target-org, it'd be good to state that explicitly in the PR description.

[nikhil-mittal-165]

Thanks for the review!

Confirmed the exit-code behavior change is intentional: bad/unauthenticated --target-org no longer fails run — ApexGuruAuthService/ApexGuruEngine in code-analyzer-core#500 catch the auth failure and skip ApexGuru with a warning (NO_ORG_CONNECTION) rather than throwing, since ApexGuru is the only consumer of target-org. I've added an explicit note about this in the PR description so it's not just implied by the diff.

On the ApexGuru analysis-mode terminal summary: agreed it's a separate concern from sanitization. Given it's small, well-tested, and low-risk, I've kept it bundled here but called it out explicitly in the description per the "one concern per PR" convention, rather than splitting it into its own PR.