yigitkonur/skills-by-yigitkonur

Use skill if you are auditing or designing a CLI for agent/LLM consumption — stable JSON output, exit codes, non-interactive flags, or iterative repair loops for machine-generated artifacts.

Use skill if you are auditing an existing MCP server for agent-readiness, designing a new MCP server before any code, or deciding framework, security, and context posture.

Use skill if you are verifying claimed-done work, auditing session/plan/branch completion, finding partial or assumed-complete tasks, or answering "are we done?" with evidence.

Use skill if you are testing or hardening an existing SKILL.md by running a fresh subagent on a real task and editing the skill where the trace shows friction.

Use skill if you are auditing a running web app's UI across pages and viewports, saving per-bug findings to a dated tree, and dispatching approval-gated fix subagents.

Use skill if you are auditing a running app's usability from real personas walking their journeys, saving per-issue findings to a dated persona/journey tree and recommending major changes, not fixes.

Use skill if you are building or auditing UI against the 30 Laws of UX — Fitts's, Hick's, Miller's, Jakob's, Gestalt, choice overload, or cognitive load.

Use skill if you are building or debugging a Chrome MV3 extension — manifest.json v3, service_worker, content_scripts, chrome.runtime/storage/alarms, popup, side_panel, declarativeNetRequest, or Web Store packaging.

Use skill if you are placing, refactoring, or auditing TypeScript mcp-use/server code under src/ and need Clean Architecture layer boundaries, folder layout, import direction, or dependency-cruiser gates.

Use skill if you are migrating email to Cloudflare Email Service, replacing Resend, SES, Postmark, or Supabase Auth mailer, wiring a send_email Worker binding, or architecting Cloudflare email sending.

Use skill if you are building TypeScript code with Effect-TS v3 — effect@3.x, @effect/*, Effect.gen, Layer, Schema, typed errors, fibers, Stream.

Use skill if you are building TypeScript browser-automation agents or apps on the Kernel SDK (`@onkernel/sdk`) — browsers, Apps, profiles, Managed Auth, pools, Playwright/CDP wiring.

Use skill if you are building TypeScript apps that import `langchain` or `@langchain/*` for agents, tool-calling, RAG retrievers, structured output, streaming, or LangGraph `StateGraph` workflows.

Use skill if you are building or maintaining a TypeScript MCP server on @modelcontextprotocol/sdk v1.x — single-package, Zod schemas, RequestHandlerExtra, McpServer.

Use skill if you are building or maintaining MCP servers on @modelcontextprotocol/server v2 alpha — split packages, McpServer, registerTool, ctx.mcpReq handler context.

Use skill if you are building or auditing TypeScript mcp-use MCPAgent code where an LLM picks and orchestrates MCP tools via run, stream, or streamEvents.

Use skill if you are writing TypeScript MCP client code with mcp-use — MCPClient, MCPSession, useMcp, McpClientProvider, mcp-use/browser, mcp-use/react, or npx mcp-use client.

Use skill if you are building TypeScript MCP servers with mcp-use/server — server.tool, response helpers, ctx.auth, sessions, transports, widgets, Inspector, deploy.

Use skill if you are authoring or fixing a Raycast Script Command (.sh or .py with @raycast.* metadata header) and need correct fields, modes, arguments, or discovery.

Use skill if you are creating a new Claude skill, redesigning one, or merging multiple skills, and need evidence-based research and comparison before writing SKILL.md.

Use skill if you are building or extending a TinaCMS-backed Next.js App Router site with tina/config.ts, MDX/git-backed content, schema modeling, useTina visual editing, or TinaCloud/self-hosted deploys.

Use skill if you are porting an existing MCP TypeScript server from @modelcontextprotocol/sdk v1.x to the v2 split-package SDK with package renames, ServerContext, and Zod v4.

Use skill if you are rebuilding a deployed site as a Next.js project AS-IS pixel-faithful from a live URL, lost-frontend repo recovery, or .html snapshot.

Use skill if you are chasing a reproducible runtime bug, intermittent failure, or repeated failed fixes and need a language-agnostic 4-phase root-cause workflow before editing product code.

Use skill if you are deploying or updating a docker-compose service on Coolify Cloud (app.coolify.io) via its API — setting domains, connecting services, env secrets, or verifying a deploy.

Use skill if you are creating, auditing, or migrating CLAUDE.md, AGENTS.md, or REVIEW.md instruction files, folder-scoped agent guidance, or cross-agent companion entrypoints.

Use skill if you are onboarding a repo to Jean or authoring jean.json and .worktreeinclude — worktree setup/run/teardown/ports proven in a throwaway test worktree, then documented in AGENTS.md.

Use skill if you are scaffolding or refreshing Makefile targets for local dev, tunnels, deploys, R2 sync, Supabase, Railway, Vercel, or Mac shipping.

Use skill if you are controlling, automating, or testing a connected Android phone via the mobilerun CLI — tapping, typing, swiping, reading the screen, or running a multi-step on-device task.

Use skill if you are publishing to npm via GitHub Actions release workflow with trusted publishing, NPM_TOKEN, provenance, semantic-release, changesets, release-please, or fixing npm publish CI.

Use skill if you are driving the agent-browser CLI for ad hoc browser automation — @ref snapshots, sessions, forms, extraction, screenshots, headed/stealth, or provider runs.

Use skill if you are driving a large, ambiguous initiative (modernize, rewrite, migrate, greenfield, or a multi-component bug hunt) and want scored multi-round question alignment before orchestrating parallel execution.

Use skill if you are fanning out many parallel Codex adversarial reviews across derived lenses, independently verifying every finding, fixing confirmed issues in grouped worktrees, and looping until clean.

Use skill if you are running Codex review loops across current branches, multiple lenses, comparisons, convergence rounds, or saved review runs.

Use skill if you are running deep multi-file research over 5+ entities or a market/category — a wave-dispatched evidence corpus on disk, executed by Claude subagents or codex exec.

Use skill if you are discovering GitHub repos, shortlisting candidates for a concrete need, or comparing OSS projects with adaptive search and repo evidence.

Use skill if you are running railway CLI commands for deploys, logs, env vars, link, ssh, db shells, scaling, or reconciling installed CLI vs upstream docs.

Use skill if you are finishing a project and want every live branch and worktree reviewed and merged into main locally, dangling branches retired, and junk swept aside.

Use skill if you are answering one technical research question with current web evidence, Reddit practitioner experience, and source-backed synthesis, optionally fanned out across parallel subagents.

Use skill if you are reviewing a PR or diff, asking for review on your branch, triaging received review feedback, or delegating review to codex.

Use skill if you are exposing a local HTTP server at a public `.ts.net` URL via Tailscale Funnel for agent-browser navigation, mobile testing, webhooks, or shared dev demos.

Use skill if you are running many small Codex-native web searches through codex exec with per-question files and parseable answer artifacts.

Use skill if you are driving the mcpc CLI (0.2.x) to test, debug, or smoke-check an MCP server over stdio or Streamable HTTP.

Use skill if you are auditing AGENTS.md/CLAUDE.md/REVIEW.md hierarchies for drift after refactors: stale file:line refs, frequency-table recounts, invalidated rules, or missing AGENTS.md for code-bearing gap folders.

Live automation — agent-browser CLI, Android device control, Tailscale Funnel public tunnels.

App & framework builders — Chrome MV3, Cloudflare Email Service, Effect-TS v3, Kernel SDK, LangChain.js, Raycast, TinaCMS+Next.js.

Instruction & config files — AGENTS.md/CLAUDE.md/REVIEW.md hierarchies, drift audits, scenario Makefiles.

Aligned delivery — scored multi-round question alignment, a filename-state spec corpus, and waved parallel subagent orchestration for a large or ambiguous initiative.

The whole pack — all 44 skills plus the internet-researcher agents. Heaviest context cost; prefer a themed bundle or single skill.

Frontend rebuild & audit — pixel-faithful URL→Next.js, UI/UX/Laws-of-UX audits.

MCP & agent interfaces — build (SDK v1/v2, mcp-use), clean architecture, convert v1→v2, test, and agent-readiness audits for MCP servers and CLIs.

Ops & release — Railway, Coolify Cloud compose deploys, repo cleanup, npm publishing.

Research & discovery — single-question and wave-based corpus research, GitHub repo scouting, bulk Codex search. Bundles the internet-researcher subagents.

Internet-researcher subagents only (no skills) — api-docs, debug-stuck, tech-choice, shipping-pattern, quick, generic. Fan them out for source-backed answers.

Review & completion — code review, review-feedback triage, done-claim audits, runtime debugging.

Skill authoring — research-driven skill creation and derailment stress-testing of existing SKILL.md files.