Plane Cloud: workspace/project AI instructions and personalized prompt settings

We need configurable AI instructions/settings in Plane Cloud.

I could not find an existing Cloud setting for this, and I could not find public docs describing workspace- or project-level AI instructions.

Requested capability:

1. Workspace-level AI instructions / system prompt.
2. Project-level AI instructions that inherit from workspace settings and optionally override/extend them.
3. Optional view-level or module-level instructions for narrower workflows.
4. Per-user personalized instructions/preferences layered on top.
5. Clear precedence order, for example:
   workspace -> project -> view/module -> user
6. A way to preview the effective instructions that an AI action or agent will actually receive.
7. API support for reading/updating these settings so automations can manage them.

Why this matters:

• teams want AI behavior aligned with their operating model, tone, taxonomy, and safety rules
• company-specific instructions like naming conventions, triage rules, decision logging rules, and escalation style should not need to be repeated in every prompt
• different projects need different behavior, for example product planning vs ops vs marketing
• individual users often need personalized defaults layered on top of team guidance

Real-world examples:

• "Use our issue taxonomy and never invent new labels"
• "When summarizing work, use board-facing language, not implementation detail"
• "For this project, prioritize blockers and dependencies before anything else"
• "For this user, default to short summaries and Croatian locale/date conventions"

Why Cloud support matters:

• this should live in the product, not in browser hacks, copy-pasted prompts, or external wrappers
• teams using Plane Cloud need a durable settings surface they can rely on across seats and workflows

This would make Plane's AI features much more usable in real organizations, especially where prompt consistency, team norms, and controlled automation matter.

No files, tests, or entry points are named. Start by locating Plane Cloud's existing AI settings and configuration surfaces, then map how workspace, project, view/module, and user scopes could be stored, inherited, previewed, and exposed through the API; done means the precedence and management experience are agreed and implemented.