dotfiles

Implement work from a ticket, spec, or inline description; auto-detects the next item from the task folders when no argument is given.

Show status overview and next steps for an .aiwork/ spec. Use when user says "next steps", "what's left", "status", "progress", or wants to see what's planned, unplanned, or incomplete for a task.

Scan current directory (recursively) for Claude Code skills, slash commands, agents, hooks, and plugin/MCP configs, then render an interactive HTML graph of their explicit and implicit relationships. Use when the user says "skills map", "skill graph", "visualize skills", "map skills", "how do my skills relate", or wants to audit a skill collection before publishing.

Turn the current conversation context into a spec. Use when user wants to create a spec (or PRD) from the current context.

Break a plan, spec, or the current conversation into vertical-slice tickets, each declaring its blocking edges.

Structure AI work artifacts under .aiwork/{YYYY-MM-DD}_{slug}/ folders using the project's protocol (triage, research, spec, prd, plan, review, notes, docs/). Use when the user asks to plan, make a spec, triage, research a task, write a report, code review, summarize findings, save the work, document decisions, record this, write that down, persist findings, or references any path under .aiwork/. Also use when starting a non-trivial task that needs a plan or spec before implementation, or when finishing one and capturing notes.

Test-driven development. Use when the user wants to build features or fix bugs test-first, mentions "red-green-refactor", or wants integration tests.

Produce a rich, interactive HTML explanation of a code change — diff, working tree, branch, or PR. Use when the user says "/explain-diff-html", "explain this diff", "explain this change", "explain this PR/branch".

Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (GLOSSARY.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions.

Start a `gv --collect` review session — open the git-history-viewer in the browser so the user authors inline review comments on the current repo, then act on the feedback they finish with. Use when the user says "gv", "gv collect", "collect review", "start a review session", "let me review", "gv review", or wants to hand-author review comments in the browser for Claude to implement.

Audit auto-memory files for the current repo (or all projects) and suggest promoting each one to a more persistent layer — global/project CLAUDE.md, an existing or new skill, a settings.json hook, or a checked-in reference doc. Use when the user says "audit memory", "promote memory", "clean up memories", "memory-promote", or wants to move ephemeral memory into durable config.

Assess a Nuxt repo against the desired tooling baseline (Vite+/pnpm/Nixpacks) and migrate it there, committing each change as it lands. Use when the user wants to bring a Nuxt project up to standard tooling, audit its setup, modernize deps/toolchain, or says "nuxt baseline", "assess nuxt", "migrate to baseline".

Fork the current Claude session into a new kitty split (new session ID), leaving this session untouched. Use when the user says "split", "branch this", "fork this session", "open this in a new window", or wants to continue a parallel conversation in a separate kitty pane. Kitty-only.

Review a PR, ticket, or the current branch in an isolated worktree. Accepts a PR id or URL, a ticket number, or nothing (current branch). Optional `--skill <name>` picks the review skill to run.

Open a real (headed) browser, drive it to a specific state in a web app — a feature, a flow, a particular screen, or a bug — then leave the window open for the user to click around in, with a short note on what to try and what to watch for. Use when the user wants to "show me X", "take me to that state/screen", "let me click around in it", "open a browser I can play with", "set up the repro", "drive me to <feature/flow/situation>", or describes any app state they want to experience firsthand.

Re-emit your own last answer tighter, following the project and global CLAUDE.md and output-style rules more strictly. Use when the user says "/tldr", "tldr", "tighten that", "too long", or "just the point".

Turn complex or visual agent responses into rich, reviewable HTML artifacts the user can annotate and send feedback on, using the lavish-axi CLI. Use when about to give a plan, comparison, diagram, table, code diff, report, or anything easier to grasp visually than as prose.

Open the last mentioned file in neovim inside a new kitty window pane. Use when user says "kt", "open in term", "open in kitty", "open in neovim", "open in nvim", "open neovim in kitty", or wants to edit a file in a terminal split. Also use proactively after generating or editing text files the user will want to review in their editor.

Compute exact dates, weekdays, working days, Czech state holidays, DST transitions, and duration sums via the bundled `datetime.ts` Deno CLI. Use whenever the answer depends on a real calendar fact (today's date/weekday in Europe/Prague, holiday lookup, working-day count, Easter-derived dates, DST direction) or on summing hours/minutes (worklogs, time tallies). Triggers: "what day is X", "is X a working day", "working days in May", "when is Velikonoce / Velký pátek", "next Tuesday", "month-end date", "DST switch", any worklog/vacation calc or hour/minute sum. Do NOT guess or do mental math.

Check Anthropic API spend and token usage from the org admin API with curl + jq. User-invoked only (/api-usage).

Transcribe audio/video files to text using ElevenLabs Scribe API. Use when user wants to transcribe audio (ogg, mp3, wav, m4a, flac, webm) or video files (mp4, mkv, mov, avi) or voice messages. Trigger on keywords like "transcribe", "voice message", "audio to text", or when an audio/video file path is provided.

Save a quick note to ~/notes/inbox as a markdown file. Use when user says "/note", "save note", "jot this down", "add to inbox".

Analyze outdated npm/pnpm dependencies and provide upgrade guidance based on release notes.

Identify and fix typos and grammar mistakes in a file, limited to syntax errors only. Use when user asks to find or fix typos, grammar mistakes, spelling, or syntax errors in a file. Does not suggest editorial or style changes.

Review a technical specification for gaps, confusing explanations, and contradictory specifications. Use when user asks to review, critique, or audit a tech spec, design doc, PRD, or RFC for high-level clarity and consistency. Does not suggest implementation details.

Use when writing or editing technical prose — tech specs, design docs, programmer articles, developer docs, RFCs, ADRs, incident writeups. Do NOT use for code comments, commit messages, or casual chat.

Czech typography rules for web development. Use when building web pages in Czech language, writing Czech content, or fixing typography issues in Czech text. Covers punctuation, dashes, quotes, numbers, units, non-breaking spaces, and HTML entities.

Extract text from scanned PDF documents and convert to clean markdown. Use when user asks to transcribe, extract text, OCR, or convert a PDF (especially scanned documents, historical documents, or image-based PDFs) to markdown format. Outputs markdown file saved next to the original PDF.

This skill should be used to generate a system health and status report for a Linux machine. Trigger when user asks for system info, system status, health check, machine report, PC report, diagnostics, or uses keywords like "sysreport", "system report", "how's my machine", "system health". Covers hardware, CPU, memory, disk, temperatures, network, Docker, top processes, listening ports, and recent errors.

This skill should be used when the user wants to extract a transcript from a YouTube video, summarize a YouTube video, or critically analyze a YouTube video's claims. Trigger on keywords like "youtube", "transcript", "summarize video", "critique video", or when a YouTube URL is provided.

Automate a real browser with the agent-browser CLI — open pages, snapshot interactive elements, click/fill by ref, extract content. Use when the user wants to browse, open a page, visit a URL, navigate a site, click or fill a form element, scrape, take a screenshot of a page, test a web app, or says "agent-browser". Works as a general-purpose browser automation tool in any repo, even ones without a local browser skill installed.

Manage Coolify deployments, resources, and servers via the coolify CLI. Use when user asks about deployments, server status, resource health, or wants to deploy/manage services on Coolify.

Open the last mentioned file in Firefox. Use when user says "ff", "open in firefox", "open in browser", or wants to preview a file in the browser. Also use proactively after generating or editing HTML/SVG/PDF files that benefit from browser preview.

This skill should be used when creating shell scripts, CLI tools, automation scripts, or any task that needs to run shell commands programmatically. Trigger when user asks to write a script, automate a task, create a CLI tool, process files in batch, or build any executable that runs shell commands. Also trigger when user mentions "dax", "deno script", or "write me a script". Deno with @david/dax is the default scripting stack on this machine - prefer it over bash scripts for anything beyond trivial one-liners.

Read the last screenshot from ~/Pictures/Screenshots. Use when user says "screenshot", "scr", "last screenshot", or references a screen capture they just took. Also use proactively when the user seems to be referencing visual content they captured.

This skill should be used when optimizing, converting, or resizing images. Trigger when user mentions image optimization, HEIC/HEIF conversion, JPEG/PNG/WebP/AVIF processing, reducing image file size, or batch image processing. Uses sharp-cli as the primary tool.