zhengxuyu

Full-stack software engineering via Claude Code CLI session with self-correction.

Review code for correctness, style, security vulnerabilities, and potential issues.

Guide product managers through evaluating whether to scale, test, or kill an acquisition channel based on unit economics (CAC, LTV, payback), customer quality (retention, NRR), and scalability (magic

Assess whether your product work is "AI-first" (using AI to automate existing tasks faster) or "AI-shaped" (fundamentally redesigning how product teams operate around AI capabilities). Use this to eva

Defines the two-axis mental model that distinguishes Director-level thinking from PM thinking: Altitude (how wide you zoom out) and Horizon (how far ahead you look). Use this to understand what actual

Diagnose overall SaaS business health by analyzing growth, retention, unit economics, and capital efficiency metrics together. Use this to identify problems early, prioritize actions by urgency, and d

Create a comprehensive company profile that extracts executive insights, product strategy, transformation initiatives, and organizational dynamics from publicly available sources. Use this to understa

Guide product managers through diagnosing whether they're doing context stuffing (jamming volume without intent) or context engineering (shaping structure for attention). Use this to identify context

Write personal blog posts with the rhythm, structure, and voice of top-tier essayists (Paul Graham, Dan Luu, Henrik Karlsson, Ben Kuhn, etc.)

Add a new feature with design, TDD, and PR workflow

Audit a codebase for security vulnerabilities, code quality issues, and architecture concerns

Use when the user says "autopilot", "auto mode", "替我回答", or "自动回答". Enables autonomous decision-making by answering Claude's own questions based on the user's personality profile.

Fix a bug with systematic debugging, TDD, and PR workflow

Create a git release — tag, push, and create GitHub release

Emergency fix — minimal change, fast PR, no design overhead

Initialize a new project — repo, structure, CI, and first commit

MVP implementation — scaffold, build modules, deploy. Zero scheduling, immediate execution. Use after locking in an idea.

Generate and filter hackathon ideas against judging criteria. Three rounds — diverge, converge, lock in. Use after selecting a competition.

Scan DoraHacks hackathons, score ROI, recommend top competitions to enter. Use when looking for hackathons to compete in.

Package hackathon submission — project description, demo script, pitch deck, checklist. Use when MVP is ready to submit.

Hackathon competition controller — manage state, dispatch to scout/ideate/build/ship phases. Use when starting or resuming a hackathon competition workflow.

This skill should be used when the user asks to "make a game", "build a game", "generate a game", or wants to generate or update a complete Godot game from a natural language description.

Execute a single Godot development task — generate scenes and/or scripts, verify visually.

Build, test, validate, and publish codemods using the Codemod platform (JSSG/ast-grep). Use when user wants to migrate frameworks, automate code transformations, or build production-grade codemods.

Autonomous Goal-directed Iteration. Apply Karpathy's autoresearch principles to ANY task. Loops autonomously — modify, verify, keep/discard, repeat. Supports bounded iteration via Iterations: N inline config.