Define repository guidance for AI coding agents

### Description

Define a concise, maintainable source of repository guidance for approved AI coding agents so agent-assisted changes follow the project’s tooling, conventions, safety boundaries, and documentation structure.

### Action Items

- [ ] Identify which AI coding agents the project intends to support and who owns the guidance.
- [ ] Decide the canonical repository instruction-file format and whether compatibility files are needed for supported tools.
- [ ] Define the durable guidance to include: workspace layout, setup and verification commands, coding/documentation conventions, PR/issue workflow, and approval boundaries.
- [ ] Link to canonical repository docs rather than duplicating volatile implementation or product information.
- [ ] Define exclusions: no secrets, personal data, unsafe blanket permissions, or instructions that override maintainers or branch protections.
- [ ] Establish review/update expectations and validate the guidance with a small, non-destructive agent-assisted task.
- [ ] Add the approved instruction file(s) only after team review.

### Additional Information

Status: Triage — refine supported tools, instruction-file strategy, scope, and ownership before implementation. This is an enablement/documentation task, not authorization for agents to make unreviewed changes.

Start by reviewing the action items and resolving which AI coding agents, instruction-file format, ownership, and scope the project will support. Define durable guidance that links to canonical repository docs and excludes unsafe permissions or sensitive information. Done means the team has reviewed the guidance, validated it with a small non-destructive task, and approved the instruction file(s).

Written by the indexing model from the issue text.