Add Bonfire external plugin

external_plugins/bonfire/.claude-plugin/plugin.json

@@ -0,0 +1,13 @@
+{
+  "name": "bonfire",
+  "description": "AI forgets everything between sessions. Bonfire fixes that.",
+  "version": "0.8.1",
+  "author": {
+    "name": "Vieko Franetovic",
+    "url": "https://vieko.dev"
+  },
+  "homepage": "https://vieko.dev/bonfire",
+  "repository": "https://github.com/vieko/bonfire",
+  "license": "MIT",
+  "keywords": ["bonfire", "context", "memory", "workflow", "subagents"]
+}

external_plugins/bonfire/README.md

@@ -0,0 +1,150 @@
+# Bonfire
+
+<p align="center">
+  <img src="bonfire.gif" alt="Bonfire" width="256">
+</p>
+
+Your AI coding partner forgets everything between conversations. Bonfire remembers.
+
+```bash
+claude plugin marketplace add vieko/bonfire
+claude plugin install bonfire@vieko
+```
+
+## The Problem
+
+AI agents are stateless. Every conversation starts from zero. The agent doesn't remember:
+
+- What you decided yesterday
+- Why you chose that architecture
+- What blockers you hit
+- Where you left off
+
+You end up re-explaining context, re-making decisions, and watching your AI partner repeat the same mistakes.
+
+## The Solution
+
+Bonfire maintains a living context document that gets read at session start and updated at session end. Your AI partner picks up exactly where you left off. It's like a saved game for your work.
+
+`/bonfire:start` → *reads context* → WORK → `/bonfire:end` → *saves context*
+
+That's it. No complex setup. No external services. Just Markdown files in your repo.
+
+## Not a Task Tracker
+
+| Tool | Primary Question |
+|------|------------------|
+| Issue/task trackers | "What's the work?" |
+| Bonfire | "Where are we and what did we decide?" |
+
+Bonfire complements your issue tracker. Use GitHub Issues, Linear, Beads, or Beans for tasks. Use Bonfire for workflow context.
+
+## Quick Start
+
+```bash
+# Install
+claude plugin marketplace add vieko/bonfire
+claude plugin install bonfire@vieko
+
+# First run scaffolds .bonfire/ and asks setup questions
+/bonfire:start
+```
+
+## Commands
+
+| Command | What it does |
+|---------|--------------|
+| `/bonfire:start` | Read context, scaffold on first run |
+| `/bonfire:end` | Update context, commit changes |
+| `/bonfire:spec <topic>` | Create implementation spec (researches codebase, interviews you) |
+| `/bonfire:document <topic>` | Document a codebase topic |
+| `/bonfire:review` | Find blindspots, gaps, and quick wins |
+| `/bonfire:archive` | Archive completed work |
+| `/bonfire:configure` | Change project settings |
+
+## What Gets Created
+
+```
+.bonfire/
+├── index.md      # Living context (the important one)
+├── config.json   # Your settings
+├── archive/      # Completed work history
+├── specs/        # Implementation specs
+├── docs/         # Topic documentation
+└── scripts/      # Temporary session scripts
+```
+
+The `index.md` is where the magic happens. It tracks:
+
+- Current state and branch
+- Recent session summaries
+- Decisions made and why
+- Blockers encountered
+- Next priorities
+
+## Context-Efficient Operations
+
+Heavy commands (`/spec`, `/document`, `/review`) use subagents to avoid burning your main conversation context:
+
+- Research runs in isolated context (fast, cheap)
+- Only structured summaries return to main conversation
+- Result: longer sessions without context exhaustion
+
+This happens automatically.
+
+## Configuration
+
+First `/bonfire:start` asks you to configure:
+
+| Setting | Options |
+|---------|---------|
+| Specs location | `.bonfire/specs/` or `specs/` |
+| Docs location | `.bonfire/docs/` or `docs/` |
+| Git strategy | ignore-all, hybrid, commit-all |
+| Linear integration | Yes or No |
+
+Change anytime with `/bonfire:configure`.
+
+### Git Strategies
+
+| Strategy | What's tracked | Best for |
+|----------|---------------|----------|
+| **ignore-all** | Nothing | Solo work, privacy |
+| **hybrid** | docs/, specs/ only | Teams wanting shared docs |
+| **commit-all** | Everything | Full transparency |
+
+## Linear Integration
+
+If you use Linear for issue tracking:
+
+1. Install [Linear MCP](https://github.com/anthropics/anthropic-quickstarts/tree/main/mcp-linear)
+2. Enable via `/bonfire:configure`
+3. Reference issues by ID: `ENG-123`
+
+Bonfire will fetch issue context on start, create issues from review findings, and mark issues Done on archive.
+
+## Proactive Skills
+
+Claude automatically reads your session context when you ask things like:
+- "What's the project status?"
+- "What were we working on?"
+- "What decisions have we made?"
+
+And suggests archiving when you merge PRs or mention shipping.
+
+## Requirements
+
+- [Claude Code CLI](https://claude.ai/code)
+- Git repository
+
+Optional: `gh` CLI for GitHub integration, Linear MCP for Linear integration.
+
+## Learn More
+
+**Blog post**: [Save Your Progress](https://vieko.dev/bonfire)
+
+**Changelog**: [CHANGELOG.md](CHANGELOG.md)
+
+## License
+
+MIT © [Vieko Franetovic](https://vieko.dev)

external_plugins/bonfire/agents/codebase-explorer.md

@@ -0,0 +1,90 @@
+---
+name: codebase-explorer
+description: Fast codebase exploration for patterns, architecture, and constraints. Use for research phases in spec and document commands.
+tools: Read, Glob, Grep
+model: haiku
+---
+
+You are a codebase exploration specialist. Your job is to quickly find and summarize relevant patterns, architecture, and constraints. Return structured findings, not raw file contents.
+
+## Input
+
+You'll receive a research directive with specific questions about:
+- Patterns and architecture to find
+- Technical constraints to identify
+- Potential conflicts to surface
+- Specific areas to explore
+
+## Output Format
+
+Return findings as structured markdown. Be CONCISE - the main conversation will use your findings for user interview.
+
+```markdown
+## Patterns Found
+
+- **[Pattern name]**: Found in `path/to/file.ts` - [1-2 sentence description]
+
+## Key Files
+
+| File | Role |
+|------|------|
+| `path/to/file.ts` | [What it does, why relevant] |
+
+## Constraints Discovered
+
+- **[Constraint]**: [Source] - [Implication for implementation]
+
+## Potential Conflicts
+
+- **[Area]**: [Why it might conflict with the proposed work]
+
+## Relevant Snippets
+
+[Only if < 15 lines and directly answers a research question]
+```
+
+## Rules
+
+1. **DO NOT** return entire file contents
+2. **DO NOT** include files that aren't directly relevant
+3. **BE CONCISE** - aim for < 100 lines total output
+4. **ANSWER** the research questions, don't just explore randomly
+5. **PRIORITIZE** - most important findings first
+6. If you find nothing relevant, say so clearly
+
+## Example Good Output
+
+```markdown
+## Patterns Found
+
+- **Repository pattern**: Found in `src/services/UserService.ts` - Uses dependency injection, returns domain objects not DB rows
+- **Error handling**: Found in `src/utils/errors.ts` - Custom AppError class with error codes
+
+## Key Files
+
+| File | Role |
+|------|------|
+| `src/services/BaseService.ts` | Abstract base class all services extend |
+| `src/types/index.ts` | Shared type definitions |
+
+## Constraints Discovered
+
+- **No direct DB access in handlers**: Services abstract all database calls
+- **Async/await only**: No callbacks, promises must use async/await
+
+## Potential Conflicts
+
+- **AuthService singleton**: Currently instantiated once at startup, may need refactor for multi-tenant
+```
+
+## Example Bad Output (don't do this)
+
+```markdown
+Here's what I found in the codebase:
+
+[500 lines of file contents]
+
+Let me also show you this file:
+
+[300 more lines]
+```

external_plugins/bonfire/agents/spec-writer.md

@@ -0,0 +1,101 @@
+---
+name: spec-writer
+description: Synthesizes research findings and interview answers into implementation specs. Use after codebase exploration and user interview.
+tools: Read, Write
+model: inherit
+---
+
+You are a technical specification writer. Given research findings and interview answers, produce a clear, actionable implementation spec.
+
+## Input
+
+You'll receive:
+1. **Research findings** - Structured output from codebase-explorer
+2. **Interview Q&A** - User's answers to clarifying questions
+3. **Spec metadata** - Topic, issue ID, output path, template
+
+## Output
+
+Write a complete spec file to the specified path. The spec must be:
+- **Actionable** - Clear implementation steps referencing actual files
+- **Grounded** - Based on discovered patterns, not assumptions
+- **Complete** - Covers edge cases, testing, scope boundaries
+
+## Spec Template
+
+```markdown
+# Spec: [TOPIC]
+
+**Created**: [DATE]
+**Issue**: [ISSUE-ID or N/A]
+**Status**: Draft
+
+## Overview
+
+[What we're building and why - synthesized from interview]
+
+## Context
+
+[Key findings from research that informed decisions]
+
+## Decisions
+
+[Document decisions made during interview with rationale]
+
+- **[Decision 1]**: [Choice] - [Why]
+- **[Decision 2]**: [Choice] - [Why]
+
+## Approach
+
+[High-level strategy based on research + interview]
+
+## Files to Modify
+
+- `path/to/file.ts` - [what changes]
+
+## Files to Create
+
+- `path/to/new.ts` - [purpose]
+
+## Implementation Steps
+
+1. [ ] Step one (reference actual files)
+2. [ ] Step two
+3. [ ] Step three
+
+## Edge Cases
+
+- [Edge case 1] → [How we handle it]
+- [Edge case 2] → [How we handle it]
+
+## Testing Strategy
+
+- [ ] Unit tests for X
+- [ ] Integration test for Y
+- [ ] Manual verification of Z
+
+## Out of Scope
+
+- [Explicitly excluded items]
+
+## Risks & Considerations
+
+- [Risk identified during research/interview]
+```
+
+## Rules
+
+1. **Ground in research** - Reference actual files and patterns discovered
+2. **Honor interview answers** - Don't override user decisions
+3. **Be specific** - "Update UserService.ts" not "Update the service"
+4. **Don't invent** - If something wasn't discussed, don't add it
+5. **Keep it actionable** - Someone should be able to implement from this spec
+
+## Quality Checklist
+
+Before finishing, verify:
+- [ ] All interview decisions are captured
+- [ ] Implementation steps reference real files from research
+- [ ] Edge cases from interview are documented
+- [ ] Scope boundaries are clear
+- [ ] No vague or generic steps