Aiding contribution/feedback with AI and templates

## Feature Proposal

### Target Project:
`.github/ISSUE_TEMPLATE/` (all issue and PR templates), `.github/skills/` (AI skills)

### Background:

In Office Hours (#2990, 20 Aug 2026) we discussed issue and PR templates. The group agreed we should brainstorm how to improve them. @willosborne raised #3006 on AI verbosity and Simplified Technical English.

As mentioned in that meeting, raising this issue to discuss personas and template improvements.

### Description:

Our templates should serve four contributor types. Most use AI tooling.

### Personas and Scenarios:

| Persona | Knowledge | Scenarios |
|:--------|:----------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| **End User** | Uses the tool or reads calm.finos.org | • Report a bug without knowing internals
• Request a feature, can only describe the problem not the fix
• Ask a question or clarify unclear documentation |
| **First-Time Contributor** | Understands CALM concepts, exploring the codebase | • Has an idea, needs help understanding how to solve it
• Picks up a "good first issue", doesn't know where to start
• Small fix, wants to contribute without ceremony
• Finds a bug or gap in the code, raises an issue and offers to fix it
• Submits a large PR without prior discussion, gets extensive feedback on unknown conventions |
| **Experienced Contributor** | Knows the codebase, conventions, and build system | • Raise and fix with minimal friction
• Design change or refactor across packages, needs structured discussion
• Signal intent before starting a roadmap item or multi-PR effort
• Disagrees with an approach, wants to propose an alternative |
| **Maintainer** | Owns the roadmap, history, and standards | • Give upfront context (including previous attempts, decisions, and why they failed) so contributors avoid rework
• Define acceptance criteria before work starts
• Triage and label quickly, templates should make intent obvious
• Communicate breaking changes or deprecations
• Limit large PRs, reviewing them is time consuming
• Wants issues to be quick to read |

### Questions:

1. Are these the right personas and scenarios? What's missing?
2. Which scenarios are poorly served by current templates, and where do templates create friction?
3. Should we use other GitHub features (Discussions, Q&A) for some scenarios instead of issues?
4. Should we apply the same standard for all personas (e.g. always include implementation detail with aid of AI) or have different expectations per persona?
5. Should we promote incubation or experimentation branches for large contributions?
6. Should AI automatically suggest a rewrite of an issue to meet standards, or just direct the author to contribution guidelines and AGENTS.md?

> **Note:** Solutions are not limited to template changes. Some or all of these could be addressed with AI skills, triage automation, or process changes.

### Some Observations:
- Feature Proposal assumes author can describe implementation and testing
- No lightweight "I have an idea" path without implementation detail
- Not clear when a maintainer can step in with guidance before a contributor goes off and builds
- No automation of "needs input" label and roadmap scheduling.

### Related:
- #2990 Office Hours 2026-08-20 (agenda item 1: review templates)
- #3006 Simpler AI output, Simplified Technical English (@willosborne)

Start by reviewing the existing files under .github/ISSUE_TEMPLATE/ and .github/skills/, then read related issues #2990 and #3006. The work is done when the personas, scenarios, and preferred template, AI, GitHub feature, or process changes have been agreed with clear acceptance criteria.

Written by the indexing model from the issue text.