feat(ai-project-generator): activity logs for AI task creation & estimate updates

[parth0025]

Summary

Adds task-level activity-log entries to the AI Project Creation flow so AI-created tasks and AI-generated estimates show up in each task's Activity tab — just like manual task actions — and attributes every entry to the current user (no more generic "AlianHub AI" label).

Closes the gaps from the screenshots: AI tasks had no "created" row, AI estimates had no "updated estimate" row, and the manual magic-wand trigger logged nothing.

What changed

orchestrator.js — activity log on AI task creation

• Logs one Task_Created history row per AI-created task (message matches the manual create format).
• Task-level only — the project Activity keeps its single "created project with N sprints and M tasks" summary instead of one row per task.
• Threads userData through createTasksForSprint → background estimator.

aiTaskEstimator.js — activity log on estimate update

• persistEstimate writes a task_total_estimate history row whenever the AI sets/updates an estimate — covering all three callers: AI bulk background, manual sidebar trigger, manual-create auto-estimate.
• "from X to Y" on a re-estimate, "set to Y" on a first estimate (shared convertToDisplayFormat).
• Guard: HISTORY.UserId is a required String — an empty actor id would fail Mongoose validation and be silently swallowed, so the log is skipped rather than attempted with a blank id.

EstimatedTime/controller.js + TaskDetailRightSide.vue — manual trigger attribution

• The magic-wand button now sends the logged-in user (userName + userId); the controller forwards it as userData, so the entry is attributed to the clicker and UserId is never blank. This is what fixes the "manual trigger not logging" bug.

AIProjectGenerator/controller.js — current-user name everywhere

• Resolves the current user's display name from uid via the company roster (already loaded) instead of the hardcoded "AlianHub AI". Propagates to all AI activity logs (project / task / estimate).

Verification (deterministic, no DB/API calls)

Check	Result
One Task_Created per task, type=task, correct project/task ids	✅
Estimate history — "set to" (first) + "from→to" (re-estimate)	✅
Empty-actor guard skips safely (no doomed write)	✅
Manual trigger logs with a valid, non-blank UserId	✅
Current-user name shown in task + estimate logs; no "AlianHub AI"	✅
All touched files syntax-check	✅

Existing functionality

• Manual task/estimate flows untouched; the estimator remains fire-and-forget and never throws. History logging is best-effort (failures logged + swallowed) so it can't affect task creation or the estimate itself.

🤖 Generated with Claude Code