docs: add AGENTS.md with worktree-based workflow guidelines

[timvw]

Summary

Add AGENTS.md with comprehensive guidelines for AI coding agents working on the wt project.

Key principle: Agents should use wt to create worktrees for each task, not switch branches in the user's main repository.

Changes

• ✅ Created AGENTS.md with agent-specific workflow guidelines
• ✅ Documents worktree-based workflow (practice what we preach!)
• ✅ Includes complete workflow examples
• ✅ Covers interactive commands and PR/MR workflows
• ✅ Lists common pitfalls to avoid
• ✅ Provides project conventions (branch naming, commits)

Why This Matters

AI agents often switch branches in the main repo, which:

• ❌ Disrupts the user's workspace
• ❌ Leaves them on unexpected branches
• ❌ Can cause confusion with uncommitted changes

Using worktrees:

• ✅ Keeps user's main repo on main branch
• ✅ Each feature isolated in its own directory
• ✅ Non-disruptive to user's workflow
• ✅ Demonstrates the tool's best practices

Example Agent Workflow

# Instead of: git checkout -b feature-branch
# Agents should do:
wt create feature-branch
# Work in: ~/dev/worktrees/wt/feature-branch/

Meta Note

This PR itself was created using a worktree! 🎯

wt create docs/add-agents-guide
# Worked in the worktree, not main repo

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 7.49%. Comparing base (d9f4c26) to head (1320f6d).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@          Coverage Diff          @@
##            main      #3   +/-   ##
=====================================
  Coverage   7.49%   7.49%           
=====================================
  Files          1       1           
  Lines        387     387           
=====================================
  Hits          29      29           
  Misses       357     357           
  Partials       1       1           

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.