YoniChechik
UserPersonal Claude Code configuration (hooks, commands, agents, settings).
Categories
Indexed Skills (40)
codex
Use when Claude wants a second opinion or hands-on help from OpenAI Codex CLI — reviewing code/plans/PRs, exploring an unfamiliar codebase, or actually running/implementing something. Codex always runs with full read/write/bash/network access; the prompt tells it whether this is review-only or a real task.
contrary
Adversarial devil's advocate. Use when the user wants their idea, plan, PR, code review, architectural decision, or claim deliberately attacked — 'poke holes in X', 'tell me why I'm wrong about Y', 'red-team this', 'what am I missing'. Distinct from /ask (neutral), /plan (constructive), /brainstorm (open exploration), /review (structured code review). Use when the user explicitly wants pushback.
plan
Explore codebase and create structured implementation plan
review
Run comprehensive code review on current branch changes
one-by-one
Present a list one item at a time, every item numbered. User says 'go' to advance, answers a numbered/yes-no question to advance AND auto-cascade to the next item in the same reply, or asks a question to stay on the item. Uses the adhd-structure skill to format each item's content. Logs answers without acting until every item is resolved, and survives interruptions (subagent/monitor notifications) by logging them too and reprinting the pending question.
plan-big-feature
Decompose a large, multi-part feature or initiative into decoupled groups (epics), research each with parallel subagents (get a second opinion from the other LLM on genuinely hard architecture calls), draft MVP-scoped tickets under each group — created in the project's ticket tracker if one exists, else a single markdown plan file as fallback — and gate ALL implementation behind a full `/one-by-one` user-approval pass. Never let implementation start before that gate completes.
session-name
Assign or update a short label describing what this session is currently doing, stored in a per-session sidecar file so a status line and the workspace title can show it; where this tool has a native rename command (e.g. Claude Code's /rename inside cmux), also triggers that so the terminal title, session picker, and any companion app pick it up too
new-feature
Start a new feature: create a worktree, then either the fast pipeline (implement → PR → merge, for small changes or when explicitly asked for 'fast') or the full pipeline (plan → implement → test → review → PR → merge). Use PROACTIVELY, without being asked by name, whenever the user asks to start any new feature or substantial piece of work — 'build X', 'add X feature', 'implement X', 'create X', 'let's start working on X', 'new feature' — anything beyond a trivial one-file edit.
devops-safe-ship
DevOps playbook for risky multi-step rollouts, migrations, and cutovers — incremental single-purpose PRs, inert-but-present fallbacks for one-line rollback, self-validation in prod (Chrome walkthroughs, /healthz probes, expected-vs-observed GO/NO-GO tables, negative controls), human-gated irreversible steps, CI-guard discipline, and durable milestone memory. Invoke for any high-blast-radius change e.g. auth/issuer cutovers, signing-key revokes.
pr-create
Create a pull request. MUST be invoked for ANY request that creates a PR — including 'create a PR', 'open a PR', 'gh pr create', 'submit for review', or any equivalent phrasing. Do NOT call `gh pr create` directly; always run this skill instead.
other-llm
Get an independent second opinion or delegate work to the OTHER coding CLI — whichever model family is NOT currently active. When the active model is an Anthropic/Claude model, run Codex; when the active model is OpenAI Codex (or another OpenAI model), run Claude. Use for independent reviews, plans, unfamiliar codebases, or hands-on implementation.
ci-watcher
Run the CI watcher script for the current or specified branch. One watcher per branch, several at once per session. Use `/ci-watcher stop` for the current branch, `/ci-watcher stop <branch>` for another, `/ci-watcher stop-all` for every watcher of this session.
adhd-plan
ADHD-friendly planning mode: one short high-level summary, then a plain list of block names, then discuss each block one at a time on request
build-tests
Plan and build tests for the current feature
pr-comments
Review and triage PR comments interactively
pr-walkthrough
Step-by-step PR code walkthrough session
create-worktree
Create isolated git worktree for feature
continue-feature
Resume work on existing feature worktree
review-tests
Review tests for quality issues
adhd-structure
THIS IS THE MOST IMPORTANT SKILL!!! Plain-language format for one chunk of content, with escalating detail: 1-line summary, then a 5-line version, then a 3-part deep dive added only on request. ALWAYS use proactively — by default, not only when asked. ALWAYS use for every markdown file and ticket (Linear issues, PR descriptions, plan/review docs).
explore
Research a topic or feature from both sides: explore the existing codebase (patterns, architecture, related files, dependencies, integration points) AND search the web for common solutions, alternatives, libraries, and blog posts about the underlying problem. Use before or during planning, or whenever unfamiliar territory needs research before deciding an approach.
new-feature-fast
Fast feature development: worktree → implement → PR (no planning, no tests)
adhd
ADHD-style output: short chunks, one item at a time, plain words, user says 'go' to advance. Works for plans, explanations, reviews — any long answer.
debug
When you get a failed result/error/user says its wrong/its a bug, use this skill to systematically debug and fix the problem.
next-steps
Surface every OPEN end blocking work — next actions, open tickets, pending decisions — as a short numbered list (title, one-line description, one-line why-it's-open, short recommendation) grouped under Open Ends. Never lists what's already done. Ends by offering to walk through them one by one. Use when the user asks 'what's next', 'next steps', 'open ends', 'what decisions are left', 'what do you need from me', or /next-steps.
human-meat-proxy
Use whenever a task might require a human-only action — approving an OAuth/app install, clicking a payment or install confirmation, revealing/copying a one-time credential, solving a CAPTCHA, or anything else Shopify/GCP/etc. structurally require a real logged-in human for. Do the entire task yourself first via Chrome MCP, right up to the blocked step, and only then ask — with the browser already sitting on the exact page, and the ask trimmed to the 1-2 clicks left.
post
Run quality, review, and review-tests checks in parallel, then fix, lint and format
quality
Run quality checks, fix code quality issues, check code style, format code, or perform lint checks
gcloud-auth
Re-authenticate an expired gcloud session. Use PROACTIVELY, without being asked by name, whenever a `gcloud`/`gsutil`/`bq`/`pulumi` command fails on credentials — 'reauthentication required', 'invalid_grant', 'invalid_rapt', 'credentials not found', 'token has expired', 'Reauthentication failed', ADC/refresh-token errors, or a `pulumi` command (e.g. `pulumi whoami`, `pulumi login`, `pulumi up` against the GCS state backend) failing with `invalid_grant`/`invalid_rapt` — or whenever the user asks to re-auth, re-login, or log in to gcloud.
stall-fallback
Detect a repeating background-agent stall bug and switch to sequential work. Use PROACTIVELY, without being asked, whenever a background/async subagent task-notification in the current session reports a failure matching 'Agent stalled: no progress for Ns (stream watchdog did not recover)' or the raw log form 'stall watchdog fired after 60000ms with no progress' — trigger on the 3rd such occurrence in the current session (i.e. more than twice), or on any occurrence in a session where 2+ have already happened.
dashboard-secret
Get a secret value that only exists in a third-party dashboard (Supabase, GCP, etc.) and seed it into keyshelf. Use whenever a fix needs a credential/API key/password that has no CLI or Management-API retrieval path — resetting a DB password, revealing a service-role/secret API key, copying an OAuth client secret, etc.
human-action-needed
Use whenever a task might require a human-only action — approving an OAuth/app install, clicking a payment or install confirmation, revealing/copying a one-time credential, solving a CAPTCHA, or anything else Shopify/GCP/etc. structurally require a real logged-in human for. Do the entire task yourself first via Chrome MCP, right up to the blocked step, and only then ask — with the browser already sitting on the exact page, and the ask trimmed to the 1-2 clicks left.
create-clone
Create isolated git clone for feature
sum
Summarize current branch work grouped by actual tasks (not by commits)
sync
Merge main, commit changes, and push
cd-permanent
Change directory within the session
feature-workflow
End-to-end feature development pipeline: plan → implement → build tests → post checks → PR → summary. Invoked by /new-feature and /continue-feature. this skill should not be invoked directly but rether by the other skills.
uv
how to run python files and mangage dependencies using UV package manager (pip replacement)
notify-waiting
Ping the user with an audible chime, green iTerm2 tab color, and a 'waiting...' terminal title to grab their physical attention. Invoke this WHENEVER Claude is about to block on a manual user action that requires the user to be at their machine — e.g., right before 2FA approval prompts, manual OAuth browser confirmations, or any interactive CLI step that pauses until the user clicks/approves something.
react-native-guidelines
Conventions and patterns for React Native/Expo mobile development including component structure, styling with NativeWind, hooks, testing, and Storybook
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.