Skip to content
Back to skills

468 Index 4d89266f

ASecurity

Complete index of all TunaCode modules

  • 9 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 11, 2026
ai-agentsdebuggingdocumentation

Works with

  • cli
  • mcp

Security analysis

A100/100

Scanned October 11, 2026

npx -y skills add tools-only/X-Skills --skill 468-index_4d89266f --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of 468 Index 4d89266f?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for 468 Index 4d89266f
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/tools-only-468-index-4d89266f/badge)](https://www.skillsdirectory.com/skills/tools-only-468-index-4d89266f)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

SKILL.md
---
title: Modules Index
description: Complete index of all TunaCode modules
---

# TunaCode Modules Index

This directory contains comprehensive documentation for all modules in the TunaCode codebase.

## Documentation Format

Each module file includes:
- **title** - Module name
- **path** - Relative path from src/
- **type** - file or directory
- **depth** - Depth from src/ root
- **description** - Purpose and responsibility
- **exports** - Main classes and functions
- **seams** - Modification (M) and Extension (D) points

## Module Organization

### Overview
- **[00-overview.md](./00-overview.md)** - Complete package structure and design principles

### Core Modules
- **[core-state.md](./core-state.md)** - Central session state management (StateManager, SessionState)
- **[core-agents.md](./core-agents.md)** - AI agent orchestration (process_request, RequestOrchestrator)
- **[core-prompting.md](./core-prompting.md)** - System prompt composition (PromptingEngine, SectionLoader)
- **[core-compaction.md](./core-compaction.md)** - Context window management (prune_old_tool_outputs)
- **[core-logging.md](./core-logging.md)** - Unified structured logging (LogManager, LogRecord, FileHandler)
- **[core-sanitize.md](./core-sanitize.md)** - Message history cleanup (run_cleanup_loop, remove_dangling_tool_calls)
- **[core-lsp-status.md](./core-lsp-status.md)** - UI-facing LSP status facade (get_lsp_status)
- **[core-configuration.md](./core-configuration.md)** - Configuration defaults and model registry facade
- **[core-constants.md](./core-constants.md)** - UI-facing constants and theme facade
- **[core-types.md](./core-types.md)** - UI-facing type exports
- **[core-messaging.md](./core-messaging.md)** - Core message content facade (get_content)
- **[core-user-configuration.md](./core-user-configuration.md)** - User config persistence facade
- **[core-system-paths.md](./core-system-paths.md)** - Session/update helpers facade
- **[core-formatting.md](./core-formatting.md)** - Formatting helpers facade
- **[core-file-filter.md](./core-file-filter.md)** - File autocomplete filter facade

### UI Module
- **[ui-overview.md](./ui-overview.md)** - Textual TUI interface (TextualReplApp, screens, renderers)

### Tools Module
- **[tools-overview.md](./tools-overview.md)** - Tool implementations and decorators (base_tool, file_tool)

### Supporting Modules
- **[configuration.md](./configuration.md)** - User settings and model registry (load_user_config, ModelRegistry)
- **[types.md](./types.md)** - Type definitions and protocols (UserConfig, MessageHistory, ToolCallback)
- **[utils.md](./utils.md)** - Shared utilities (estimate_tokens, estimate_message_tokens, parse_json)
- **[utils-limits.md](./utils-limits.md)** - Tool output limits with cascading defaults
- **[lsp.md](./lsp.md)** - Language Server Protocol client (LSPClient, get_diagnostics)
- **[prompts.md](./prompts.md)** - Modular prompt sections (agent_role.md, critical_rules.md)
- **[exceptions.md](./exceptions.md)** - Custom exception hierarchy (TunaCodeError, ToolExecutionError)
- **[constants.md](./constants.md)** - Global constants and UI text (UI_COLORS, TOOL_NAMES)

## Quick Reference

### By Responsibility

**Agent & AI:**
- core-agents.md
- core-prompting.md
- core-state.md
- core-sanitize.md

**Logging & Debugging:**
- core-logging.md

**User Interface:**
- ui-overview.md
- constants.md

**Tools:**
- tools-overview.md

**Configuration:**
- configuration.md

**Utilities:**
- types.md
- utils.md
- lsp.md

**Error Handling:**
- exceptions.md

### By Directory

```
src/tunacode/
├── core/
│   ├── agents/ → core-agents.md
│   │   └── resume/sanitize.py → core-sanitize.md
│   ├── logging/ → core-logging.md
│   ├── prompting/ → core-prompting.md
│   ├── state.py → core-state.md
│   ├── compaction.py → core-compaction.md
│   ├── lsp_status.py → core-lsp-status.md
│   ├── configuration.py → core-configuration.md
│   ├── constants.py → core-constants.md
│   ├── types.py → core-types.md
│   ├── messaging.py → core-messaging.md
│   ├── user_configuration.py → core-user-configuration.md
│   ├── system_paths.py → core-system-paths.md
│   ├── formatting.py → core-formatting.md
│   ├── file_filter.py → core-file-filter.md
├── ui/ → ui-overview.md
├── tools/ → tools-overview.md
├── configuration/ → configuration.md
├── types/ → types.md
├── utils/ → utils.md
│   └── limits.py → utils-limits.md
├── lsp/ → lsp.md
├── prompts/ → prompts.md
├── exceptions.py → exceptions.md
└── constants.py → constants.md
```

## Seams Legend

- **M (Modification)** - Points where you can modify existing behavior
- **D (Extension)** - Points where you can extend with new functionality

## Usage

1. **For Understanding:** Start with 00-overview.md, then dive into specific modules
2. **For Development:** Refer to specific module documentation before making changes
3. **For Onboarding:** Read overview, then core modules, then supporting modules

## Generated By

This documentation was generated using the Gemini MCP tool for code analysis.
- Analysis date: 2026-01-04
- Depth: 0 (top-level modules)
- Analyzer: Gemini 2.5 Flash

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.