research: Understand /retro vs close-issue retro taxonomy

## Research Task
Investigate and document the relationship between different retro mechanisms in the Claude Code workflow.

## Current Understanding
- User mentions using "close-issue constantly" with automatic retro
- Separate "/retro" command exists but hasn't been used yet
- Both appear to be slash commands
- Close-issue includes retro sometimes but user can opt out

## Questions to Answer
1. Where is the close-issue command defined?
2. Where is the /retro command defined?
3. What triggers the automatic retro in close-issue?
4. What's the scope/purpose difference between them?
5. Is there code reuse between the two?
6. Are these Claude Code built-in or custom commands?

## Investigation Steps
- [ ] Search for close-issue implementation
- [ ] Search for /retro implementation
- [ ] Check Claude Code documentation
- [ ] Look for custom command definitions
- [ ] Document findings

## Success Criteria
- Clear understanding of both retro mechanisms
- Documentation of when to use each
- Identification of any code reuse opportunities
- Recommendations for improvements

Related: #591 (think triggers), #592 (parent task)

Search the repository for the close-issue and /retro command definitions, then check the Claude Code documentation and custom command definitions mentioned in the investigation steps. Compare their triggers, scope, purpose, and any shared implementation. Done means documenting the findings, when to use each mechanism, code-reuse opportunities, and improvement recommendations.

Written by the indexing model from the issue text.