GenesisTools

Control macOS UI via the Accessibility API and record short screen captures reviewed frame-by-frame. Use when automating native apps — clicking buttons, filling forms, reading element state, finding elements, getting window bounds — or when the user wants a short screen recording of an animation/transition/flicker reviewed via a contact sheet, optionally pushed to a vitrinka board. Triggers on "click button in app", "fill form in native app", "automate macOS app", "list UI elements", "find button", "read text field value", "get window position", "interact with native app", "record the screen", "capture this transition", "Xs recording", "Nfps", "watch this animation", "why does it jump/flicker", "record this app for a few seconds", "push the recording to a board".

Drive a real, extension-loaded Chrome/Brave browser for the GenesisTools YouTube extension via chrome-devtools-mcp as a standalone MCP client — no Claude Code config edits, no restart. Use whenever the user wants to test, debug, screenshot, or click through the YouTube extension's UI (side panel tabs, popup, content script), asks to "load the extension and click X", wants a screenshot of the extension mid-interaction, or needs to find exact pixel coordinates to click something inside a web page (not a native macOS app — for that, use the screen-capture skill instead). Also use this whenever chrome-devtools-mcp's own tools are attempted and fail with "no browser" or launch a blank vanilla browser with no extension loaded — that means the server wasn't wired to a real browser, and this skill is the fix.

Cross-agent messaging protocol via `tools agents`. Use when the main agent is about to spawn subagents that should be able to message each other (or back to the main agent) during their work. Teaches the login (auto-registers) / message / discover / listen pattern, the stream-vs-once receive modes, and the no-loss delivery contract.

Empirically evaluate and trim ANY agent-instruction file — user/project CLAUDE.md, CLAUDE.local.md, AND AGENTS.md (the cross-tool standard read by Codex, Cursor, Gemini CLI) — to find rules a clean model already knows, rules the repo/hooks already enforce, and internal duplication, then propose evidence-backed cuts. Autodetects which files exist and which the harness actually loads, runs headless claude with ALL instruction files EXCLUDED (verified by canary) against MULTIPLE models, and classifies every rule KEEP/SHRINK/CUT with a multi-provider rubric. Use when the user says "improve agents.md", "evaluate agents.md", "trim CLAUDE.md", "improve claude md", "is my CLAUDE.md bloated", or "/improve-agents-md [path]".

Create and maintain a compaction-proof progress file (.claude/plans/<plan>.handoff.md) next to an implementation plan, so any executor - fresh session, post-compact resume, or parallel subagent - knows exactly where the work stands and which plan lines to read next. Invoke when starting to execute a plan, when resuming one, when handing work to another agent, or when the user says "create a handoff" / "prepare the handoff".

Write comprehensive implementation plans (subsumes writing-plans - bite-sized TDD tasks, exact paths, complete code) hardened with an executor-proof contract - all judgment spent at plan time, REASON/VERIFY/ON-FAIL per task, no-improvisation covenant - so even the weakest model executes to strong-model-quality output. Invoke INSTEAD of writing-plans whenever the plan may be executed by another session, subagent, or weaker model.

Save/apply/unapply named code overlays across projects with `tools stash`. Use when the user wants to capture a chunk of working-tree changes for re-use, apply a previously saved stash into the current project, shape an applied stash by editing source then capturing the result (update), surgically remove an applied stash with diff review (unapply), inspect drift between stored and applied content (diff), or list/inspect stashes. Triggers on "stash this", "save this overlay", "apply my <name> stash", "pop my debug stash here", "what stashes do I have applied", "merge stashes", "curate this overlay".

Azure DevOps work items, queries, and dashboards. Use for "get workitem", "fetch task", "show query", "analyze task", or any Azure DevOps URL. For time logging (Timely sync, Clarity fill), invoke the `/gt:timelog` command instead — this skill defers to it.

Hypothesis-driven runtime debugging with temporary, auto-cleanable instrumentation across Node/TS, PHP, and browser (via HTTP). Triggers on "debug this", "add temporary logs", "instrument X", "why is this slow/wrong at runtime", "test my theory", comparing failing vs passing runs, intermittent or cross-device bugs. Not for type errors or static code review.

Use when user says "research X", "look up X online", "what's the current state of X", "what are people saying about X", "find examples of X on github", "compare options for X", "is X worth using", "what do people complain about with X", or otherwise asks for information from outside the local codebase.

Run interactive CLI commands (Metro, Vite, dev servers) with PTY capture and agent-friendly log tail. Use when the user wants to run a long-lived process with log capture, grep session logs, follow live output, or inspect session state.

Answer a question and preserve the Q→A for later review. Use when the user runs `/question <x>` (answer it, then save), or right after you give a substantive answer worth keeping to a question interjected mid-session ("why X over Y", "how does Y work", decisions, rationale, tradeoffs). Not for routine instructions you execute or quick status checks/acknowledgements ("ok", "thanks", "pushed yet?").

Create Timely time-log entries from auto-tracked memories. Use when the user says "log my Timely day", "convert memories to events", "fill Timely from yesterday", "create timely entry from memory", "split my day into projects", or "categorize my Timely time". Drives a non-interactive **plan / apply** workflow that lets you (the LLM) split a day's memories across multiple projects. Does NOT cover ADO TimeLog or Clarity sync — for those, defer to /gt:timelog.

Token-efficient HAR (HTTP Archive) analysis with a reference system. Use for any .har file, analyzing network traffic, debugging API calls, or reviewing HTTP requests. Never cat/jq HAR files — always use this.

Find or reference a past Claude Code conversation — fixes, decisions, or discussions from earlier sessions ("you helped me fix", "we debugged", "I remember asking"). Locate sessions by topic, file, or date. Not for codebase/git/Slack history.

Triggers when rebasing a stack of branches — a parent plus child branches on top — where rebasing the parent leaves children orphaned. Use for "cascade rebase", "rebase branch hierarchy", "update child branches after rebase", "my sub-branches broke after rebasing parent", or any scenario that needs `git rebase --onto`. Not for a single-branch rebase.

Read or search GitHub (issues, PRs, review threads, comments, code, notifications) and analyze GitHub Actions (workflow runs, billing, costs, failures, rerun/cancel). Use on any GitHub URL or repo lookup. Not for implementing code fixes or PR feedback — use gt:github-pr.

Self-maintaining docs system. Bootstraps, validates, and refines minimal codebase documentation. Use for creating, updating, or auditing project docs.

Inspect React Compiler (babel-plugin-react-compiler) output. Use for "compiler output", "see compiled", "why isn't this memoized", or comparing original vs compiled components.

Summarize a Claude Code session — extract learnings, generate postmortem/changelog, write onboarding docs, or document what was done. Use for "summarize this session", "extract learnings", "postmortem", "write up what we did". Not for general text summarization.

Fix TypeScript compile errors and eliminate `any` types. Use on failing builds, type-safety audits, or `any` cleanup.

Manage project-scoped todos with rich context capture. Use when the user wants to track tasks, create reminders, link work items to PRs/issues, or manage work across sessions. Triggers on "add todo", "create task", "remind me to", "track this", "todo list", "what's on my plate", "mark done", "complete task", "what should I work on". Also use proactively when the user mentions wanting to remember something for later or needing to follow up.