feat: add bug-investigator agent

[pokhiii]

Summary

• Adds a new bug-investigator Claude Code agent that systematically investigates bugs — from clarifying symptoms to isolating root causes to proposing solution options with trade-offs
• Designed as a first pass before implementation: human reviews findings, then implementation-planner handles the fix plan in the second pass
• Follows the same templatized, placeholder-based conventions as the existing implementation-planner agent

Agent workflow

Bug reported → bug-investigator → investigation report on GitHub issue
                                        ↓
                             Human reviews & validates
                                        ↓
                        implementation-planner → fix plan

5-step investigation process

1. Expected vs. actual behavior — clarify what should happen and what is happening
2. Reproduction profile — who, where, when, how, what data triggers it
3. Root cause isolation — trace through code to specific files, functions, and lines
4. Solution options — 2–3 approaches with complexity, risk, effort, and trade-off analysis
5. Post to GitHub issue — for team visibility and human review

Files

• claude/agents/bug-investigator/bug-investigator.md — agent template with placeholders
• claude/agents/bug-investigator/README.md — documentation with placeholder reference and Django/Next.js examples
• claude/README.md — updated Available Agents table
• README.md — updated Recipes table

Test plan

• Verify the agent template has valid YAML frontmatter (name, description, model, color, memory)
• Verify all placeholders are documented in the README with examples for both Django and Next.js
• Verify the setup prompt in claude/README.md still discovers all agents automatically
• Set up in a test project and trigger with a bug report to validate end-to-end flow

🤖 Generated with Claude Code