Compress the whole thread — the problem or feature being worked on and how it was solved — into a single TL;DR line, then propose up to three terse next-step labels (≤8 words each, slash commands welcome) the user can take. Use whenever the user types `/tldr`, `/recap`, says "tldr", "tl;dr", "summarize and suggest next steps", "recap that", "what should I do next", "give me the gist", or otherwise asks for a quick summary of what's been going on plus suggestions for what to do next. Trigger even when the user phrases it casually ("ok so what now?", "give me the short version + next steps") — this skill exists exactly for those moments where a long thread needs distilling and a clear handoff to action. Use the "Variant: dyslexia-friendly visual preview" section on `/tldr visual` or `/tldr dyslexia`, or when the user asks for a dyslexia-friendly, visual, or cmux recap.
Orchestrate a large task across multiple agents with cost-aware model routing — session model for judgment-heavy work, Sonnet for mechanical work, Fable or GPT-5.6-sol-and-newer via Codex CLI for the hardest nodes and cross-model review. Decomposes, routes, picks fan-out vs team topology, presents the plan, executes, and verifies. Use when the user invokes /dynamic-agents, says "dynamic agents" or "fan out dynamic", explicitly asks to parallelize or split work across agents or models, or hands over a task too big for one agent pass (multi-subsystem feature, repo-wide audit or migration, research plus implementation) whose subtasks are independent enough to run in parallel. NOT for tasks one agent can finish inline, even multi-file ones.
Align recently written or changed code so it matches the surrounding project's established conventions — formatting, naming, imports, comments, and idioms — learned from the codebase itself, then apply the fixes. Use this whenever the user wants their changes to "match the codebase / our conventions / the rest of the app", asks to make a diff or PR style-consistent, wants a style pass or style review before opening a PR, says code "feels off" or inconsistent with the project, or invokes /code-style. Works in any language or framework — it discovers the rules from the repo rather than assuming them. Reach for this even when the user doesn't say the word "style" but clearly wants new code to blend in with existing code. Also covers de-slopping AI-generated artifacts from the diff — over-commenting, defensive checks or try/catch abnormal for the code path, casts to any / type-ignore escape hatches, needless nesting — so use it when the user says "remove AI slop", "de-slop", "this looks AI-written", or wants AI a
Turn an agent-browser session into a polished, professional product-demo video — fully local, free, no cloud upload. Records the browser to WebM while logging every action, then composites an MP4 with automatic zoom-to-click, a smooth animated cursor with trail, click ripples, an on-screen keystroke overlay (keycast), idle-time trimming, chapter markers (both burned-in lower-thirds and embedded MP4 chapters), and a gradient background with a rounded, shadowed browser card. Use this whenever the user wants to record or generate a demo video, screencast, walkthrough, "how it works" clip, or animated screen recording of a web flow — especially one that looks like Screen Studio (auto-zoom, keystrokes, chapters) but produced by the agent itself. Trigger on "record a demo", "make a screencast", "generate a walkthrough video", "screen recording of this flow", "demo video with zooms and chapters", "record the signup/onboarding flow". Built on top of the agent-browser CLI; requires ffmpeg and Node. Not for live screen
Use when a defect's mechanism is known and the change you are about to ship may leave it intact — prompt/instruction tweaks, retries, warnings, post-hoc repair, fallbacks, guards downstream of the bad value — or when cost, risk, validation difficulty or scope is steering you toward a substitute. Fires on YOUR OWN reasoning, not just the user's words: "too risky to change", "hard to validate", "we detect and correct it downstream", "the prompt now instructs it to", "out of scope for this fix", "we don't have that information", "that's decided by the model". ALWAYS load before writing "fixed" or "root cause" in any summary, commit or PR. Not for initial diagnosis (use systematic-debugging), nor for routine fixes that plainly remove the mechanism.
Analyze recent Claude Code sessions and report where the most tokens (and estimated cost) were burned, with data-driven tips to cut usage and avoid rate limits. Scans ~/.claude/projects session logs over a window (default last 5h, configurable), ranks the heaviest sessions and projects, breaks down cache efficiency and model spend, and renders a minimalist report — in a cmux markdown panel when cmux is available, otherwise as an ASCII summary in the terminal. Use this whenever the user wants to know where their tokens went, why they're hitting (or nearing) usage/rate limits, what their recent Claude Code activity cost, which sessions or projects are the most expensive, or how to optimize token usage. Trigger on phrases like "token burn", "where did my tokens go", "token/usage report", "what's burning tokens", "session cost", "how much have I spent on Claude", "why am I hitting rate limits", "analyze my claude sessions", "token breakdown", "/token-burn", and "/burn".
Packages the current chat (or one or more specific issues/topics) into self-contained handoff files and launches each in its own fresh `claude` session with a dedicated, meaningfully named git worktree — so work continues without manual context copy-paste. Each child claude inherits the CALLER session's permission mode (no hardcoded bypass). Requires cmux: by default each handoff opens as a new cmux workspace; with `--tabs` (or "in this/the same workspace", "as tabs") they open as tabs in the current workspace. Use when the user wants to hand off work to a fresh Claude in its own worktree instead of continuing inline. Trigger phrases: "handoff to worktree", "handoff-to-worktree", "spin this off into its own claude", "fork this into a worktree", "run it in a separate claude". No argument = handoff the whole chat (one worktree); argument(s) = one handoff/worktree per issue or topic.
Hands-free voice loop for driving Claude by voice. Combines the say skill (Gemini TTS audio out), a focus-pull back to the caller's cmux prompt, Wispr Flow dictation (voice in), and an appendix-stop listener so the user ENDS and SUBMITS each turn by saying the word appendix. When active, every reply is spoken aloud and the mic is auto-armed, so the user answers and sends without touching the keyboard. Use when the user asks for hands-free mode, to control Claude by voice, to start the voice flow, or voice mode on. Turn it off with hands-free off, stop voice mode, or stop the voice flow.