Skill Claude CodeCodex
Use whenever designing, reviewing, refactoring, or evolving the STRUCTURE of a codebase — module boundaries, layering, where new code should live, naming a package or service, an architecture-worthy decision, or auditing existing structure. Triggers on phrases like "design the architecture", "where should this live"…
Skill Claude CodeCodex
Fork the current Claude Code conversation into a sibling Apple Terminal tab (default) or window. Use when the user invokes /claude-forkchat, /claude-forkchat tab, /claude-forkchat window, or asks to "fork this chat", "branch this conversation into another tab/window", "open a fork alongside", "clone this session"…
Skill Claude CodeCodex
Open a fresh (non-resumed) claude session in a new Apple Terminal tab or window, in the same working directory as the current session. Use whenever the user invokes /claude-samefolder, /claude-samefolder tab, /claude-samefolder window, or asks to "open another claude in this folder", "start a fresh claude alongside…
Skill Claude CodeCodex
Set up Claude Code's status line to show, left to right, the percentage of the context window used, the conversation cost in USD, a compact account+usage token (the signed-in Anthropic account's @name joined with the 5-hour and 7-day plan-usage percentages, e.g. @example12%40%), a short model name carrying that…
Skill Claude CodeCodex
Continue the CURRENT Claude Code conversation IN PLACE — taking over the current Apple Terminal tab — with the bypass-permissions mode flipped. Use when the user invokes /claude-toggle-permissions or asks to "switch this session to bypass mode", "turn off dangerously-skip-permissions", "toggle permissions", "continue…
Skill Claude CodeCodex
Progressively teach the user about a topic — a new codebase, a library or system, or an abstract concept — starting at the highest level and descending one controlled layer at a time, calibrated to what they already know so they are never overwhelmed and never lost. Use this whenever the user invokes /dig-into, or…
Skill Claude CodeCodex
A categorized, sourced reference catalog of every AI product, model, and feature Google announced at Google I/O 2026 (May 19–20, 2026), plus the pre-existing products they build on. Use this whenever the user asks "what is ", "which Google AI tool should I use for ", "what did Google announce at I/O 2026", "what's the…
Skill Claude CodeCodex
Schedule persistent local background jobs on macOS via launchd — plain shell commands or agentic Claude prompts run headless on a cron schedule, surviving reboots and catching up after sleep. Use whenever the user invokes /schedule-local, or asks to "schedule a job", "run this every morning", "cron this command", "set…
Skill Claude CodeCodex
Design and build Slack integrations — bots, apps, incoming webhooks, slash commands, Block Kit UI, Workflow Builder steps, and AI agents that live in Slack. Use this whenever work touches Slack's platform in any way — choosing between a webhook, a token script, a no-code workflow, a custom workflow step, a Socket Mode…
Skill Claude CodeCodex
Drive an Android emulator, an iOS simulator, or a Chrome browser from the command line — reading the UI as structured text (accessibility tree or DOM) and acting on elements by name, instead of burning tokens on screenshots and guessing tap coordinates. Use this whenever the work involves controlling a phone…
Skill Claude CodeCodex
Audit a website codebase against Google's Search Quality Rater Guidelines (E-E-A-T, YMYL, page purpose, content structure, attribution, spam patterns, trust signals). Use whenever the user invokes /web-audit-with-google-guidelines, asks to "review the site for Google quality", "check E-E-A-T", "audit for SEO quality…
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: