ITB AI Tools
A collection of AI coding and productivity tools — skills, agents, hooks, and more — for Claude Code, Codex, and other AI coding assistants.
Installation
Claude Code
Add this marketplace to Claude Code:
/plugin marketplace add it-bens/ai-tools
Then browse and install plugins:
/plugin
Codex
Add the repository as a Codex plugin marketplace:
codex plugin marketplace add it-bens/ai-tools
Then start Codex, open /plugins, and install any plugin marked as compatible below.
Available Plugins
| Plugin | Description | Category | Codex compatibility |
|---|
| behavior-diagnostics | Debug AI tooling behavior with root cause analysis and honest self-diagnosis across sessions | Development | ❌ |
| clipboard-copy | Bash-based MCP server with cross-platform clipboard tools — clipboard_copy for inline text and clipboard_copy_file for an absolute file path; auto-detects pbcopy / wl-copy / xclip / xsel / clip.exe with an OSC 52 fallback, plus shared Claude Code/Codex hooks that prime the agent with MCP tool directives and block native clipboard-write Bash commands | Productivity | ✅ |
| code-comment-writer | Code-comment-writing skill that improves comments toward 'why not what': removes redundant, improves vague, condenses verbose, preserves valuable, and flags inconsistent comments across files, directories, or git changes. Deterministic scope resolution via scope.sh, an uncertainty pass for risky edits, a read-only mode, and Pre-Step-N / Post-Step-N extension with named values | Development | ❌ |
| code-comment-writer-extension-setup | One-time setup skill that provisions the overlay content file and the PostToolUse:Skill + UserPromptSubmit hook entries needed to extend the code-comment-writer plugin. Long plugin and skill names prevent accidental invocation | Development | ❌ |
| codex-integration | Claude Code integration that consults OpenAI Codex for fresh perspective — auto-escalation when stuck, on-demand second opinions, and web research | Development | ❌ |
| commit-message-writer | Commit-message-writing skill for Conventional Commits. Mode detection (staged / squash / rewrite), a deterministic gather-script-driven workflow, anti-slop validation via the human-author:ai-slop-writing-fixer subagent (auto-installed by Claude Code and packaged separately for Codex), and validation mode. Extendable at every step via Pre-Step-N / Post-Step-N positions and via recognized named configuration values | Development | ✅ (custom agent required) |
| commit-message-writer-extension-setup | One-time setup skill that provisions the shared overlay content and host-specific delivery configuration needed to extend the commit-message-writer plugin in Claude Code or Codex. Long plugin and skill names prevent accidental invocation | Development | ✅ |
| explore-with-opus-enforcer | Enforces Opus model for Explore subagent for maximum exploration accuracy | Guardrails | ❌ |
| explore-with-sonnet-enforcer | Enforces Sonnet model for Explore subagent to prevent lossy Haiku summaries | Guardrails | ❌ |
| file-read | Portable MCP read_file tool for local text files with file_path, offset, and limit arguments, returning line-numbered output so coding agents can avoid shell readers such as cat, head, tail, or sed -n | Development | ✅ |
| human-author | Author content that reads like a human wrote it — currently ships the ai-slop-writing-fixer subagent that scrubs LLM-pattern violations and returns the fixed prose with a structured change report | Productivity | ❌ |
| llm-author | Author LLM-targeted content — prompts, skills, agents, rules files, plus session handoff and feedback prompts (model-invoked on explicit request, crafted via prompt-engineering) — for Claude 4, GLM 4.7, and Gemini 3 | Productivity | ✅ |
| native-tools-enforcer | Enforces native search tools (Grep/Glob on classic builds, bfs/ugrep on native macOS/Linux) via PreToolUse hook; SessionStart hook primes Claude with mode-appropriate directives; setup skill installs binaries and configures mode | Guardrails | ❌ |
| plan-with-opus-enforcer | Enforces Opus model for Plan subagent for maximum reasoning depth | Guardrails | ❌ |
| plan-with-sonnet-enforcer | Enforces Sonnet model for Plan subagent to ensure thorough architectural reasoning | Guardrails | ❌ |
'%20stop-opacity%3D'0.16'%2F%3E%3Cstop%20offset%3D'1'%20stop-color%3D'rgb(200%2C90%2C60)'%20stop-opacity%3D'0.03'%2F%3E%3C%2FlinearGradient%3E%3C%2Fdefs%3E%3Crect%20width%3D'320'%20height%3D'200'%20fill%3D'url(%23g)'%2F%3E%3Ccircle%20cx%3D'250'%20cy%3D'56'%20r%3D'92'%20fill%3D'rgb(200%2C90%2C60)'%20fill-opacity%3D'0.06'%2F%3E%3Ccircle%20cx%3D'64'%20cy%3D'172'%20r%3D'58'%20fill%3D'rgb(200%2C90%2C60)'%20fill-opacity%3D'0.05'%2F%3E%3C%2Fsvg%3E)
