thoughtbox-dot-claude

A comprehensive, reusable .claude/ configuration folder for Claude Code CLI. This repository contains a curated collection of skills, commands, hooks, plugins, agents, and a learning memory system to enhance Claude Code workflows.

Quick Start

Copy the .claude/ folder into any project:

# Clone and copy
git clone https://github.com/Kastalien-Research/thoughtbox-dot-claude.git
cp -r thoughtbox-dot-claude/.claude/ /path/to/your/project/

What's Included

`/agents` - Custom Sub-Agents

Specialized agents for specific tasks:

meta-agent.md - Generates new sub-agent configuration files from descriptions
fact-checking-agent.md - Verifies claims in documentation against sources of truth
reasoning-evaluator.md - Evaluates reasoning quality and logical consistency

`/ai_docs` - Reference Documentation

Curated documentation for AI context:

Anthropic docs - Claude Code hooks, subagents, custom commands, output styles
OpenAI quick start - Reference for OpenAI API patterns
Memory system design - Full design documentation for the memory system
uv single-file scripts - Python scripting patterns

`/bin` - Memory CLI Tools

Command-line utilities for the memory system:

Command	Description
`memory-add`	Add new learnings to memory
`memory-query`	Search memory for relevant patterns
`memory-stats`	View memory system statistics
`memory-rank`	Rank memory entries by relevance
`memory-pipe`	Pipe content into memory

`/commands` - Slash Commands (Skills)

Organized collection of 60+ slash commands:

Core Categories

Category	Description	Key Commands
analysis/	Code analysis & insights	`/context-aware-review`, `/knowledge-graph`, `/evolution-tracker`
clear-thought/	Structured thinking	`/think-solve`, `/think-debug`, `/think-decide`, `/think-create`
debugging/	Systematic debugging	`/systematic-debug`, `/distributed-debug`, `/temporal-debug`
development/	Code generation	`/implementation-variants`, `/spec-to-test`, `/implement-spec`
exploration/	Solution discovery	`/parallel-explorer`
loops/	OODA loop building blocks	Authoring, exploration, verification, refinement loops
meta/	Higher-order frameworks	`/ulysses-protocol`, `/virgil-protocol`, `/learning-accelerator`
orchestration/	Multi-agent coordination	`/swarm-intelligence`, `/meta-orchestrator`, `/adaptive-workflow`
research/	Research workflows	`/docs-researcher`
speckit.*	Specification workflows	`/speckit.specify`, `/speckit.plan`, `/speckit.tasks`, `/speckit.implement`
startup/	Session initialization	`/prime`, `/metacursor`, `/metaclaude`
synthesis/	Pattern extraction	`/pattern-synthesizer`, `/knowledge-fusion`, `/wisdom-distillation`
toolscontext/	MCP tool context	`/exa`, `/firecrawl`, `/context7`
workflows/	MCP orchestration	`/mcp-workflow`, `/mcp-chain`, `/mcp-orchestrate`

Memory Commands

/memory-start - Initialize memory system
/memory-add-quick - Quick add to memory
/memory-search - Search memory

`/hooks` - Lifecycle Hooks

Event-driven shell scripts for Claude Code:

Hook	Trigger	Purpose
`pre_tool_use.sh`	Before tool execution	Block dangerous git operations, validate commands
`post_tool_use.sh`	After tool execution	Log operations, track file access
`session_start.sh`	Session begins	Load context, show memory status
`session_end_memory.sh`	Session ends	Prompt for learning capture
`user_prompt_submit.sh`	User submits prompt	Pre-process user input
`git-validator.sh`	Permission requests	Advanced git validation (block/approve/prompt)
`notification.sh`	Various events	Desktop notifications
`track_file_access.sh`	File operations	Track which files agents access
`memory_pattern_detector.sh`	Periodic	Identify memory gaps and patterns

Protection Features:

Blocks force pushes and direct pushes to protected branches
Prevents dangerous rm -rf operations
Validates commit message format
Logs all git operations for audit trail

`/hooks-alternate` - Alternative Hook Set

Alternative hook implementations with additional utilities:

utils/llm/ - LLM API wrappers (Anthropic, OpenAI, Ollama)
utils/tts/ - Text-to-speech integrations (ElevenLabs, OpenAI, pyttsx3)

`/output-styles` - Response Formatting

Customize Claude's output format:

Style	Description
`ultra-concise.md`	Minimal, essential information only
`bullet-points.md`	Structured bullet point format
`markdown-focused.md`	Rich markdown formatting
`table-based.md`	Tabular data presentation
`yaml-structured.md`	YAML-formatted res

...

kastalien-research/thoughtbox-dot-claude

README

thoughtbox-dot-claude

Quick Start

What's Included

`/agents` - Custom Sub-Agents

`/ai_docs` - Reference Documentation

`/bin` - Memory CLI Tools

`/commands` - Slash Commands (Skills)

Core Categories

Memory Commands

`/hooks` - Lifecycle Hooks

`/hooks-alternate` - Alternative Hook Set

`/output-styles` - Response Formatting

Publisher

Statistics

kastalien-research/thoughtbox-dot-claude

README

thoughtbox-dot-claude

Quick Start

What's Included

/agents - Custom Sub-Agents

/ai_docs - Reference Documentation

/bin - Memory CLI Tools

/commands - Slash Commands (Skills)

Core Categories

Memory Commands

/hooks - Lifecycle Hooks

/hooks-alternate - Alternative Hook Set

/output-styles - Response Formatting

Publisher

Statistics

`/agents` - Custom Sub-Agents

`/ai_docs` - Reference Documentation

`/bin` - Memory CLI Tools

`/commands` - Slash Commands (Skills)

`/hooks` - Lifecycle Hooks

`/hooks-alternate` - Alternative Hook Set

`/output-styles` - Response Formatting