崔鹏飞 Cui Pengfei
Community@cuipengfei
cuipengfei.me
Agent Skills by 崔鹏飞 Cui Pengfei
Showing 22 vetted skills indexed across 1 GitHub repositories.
quality-standards
Enforce SOLID, DRY, and maintainability rules in code review and refactoring.
ba-collaboration
Transforms BA requirements into user stories and acceptance criteria via structured AI collaboration.
programming-workflow
Enforce a TDD red-green-refactor workflow across preparation, design, test, implementation, and review stages.
planning-workflow
Transform ambiguous project ideas into structured Markdown implementation plans.
natural-writing
Remove 24 common AI writing patterns from generated text.
jpm-voice
Enforce a human-like JPM/PJL voice with 鬼畜破腔 pressure in technical responses.
jichang
Evaluate proxy provider trustworthiness by classifying evidence tiers and risk markers.
codex-advisor
Invoke OpenAI Codex MCP for second-opinion code review with categorized feedback.
multi-source-inquiry
Cross-verify claims across multiple source categories with evidence labels.
auto-extract
Extracts learning signals from chat sessions and persists them to configuration files.
knowledge-fetch
Retrieve historical knowledge from four layered memory stores with grouped results.
improve-prompt
Rewrite user prompts into structured instructions with explicit goals and constraints.
shortcut-system
Parse `!`-suffixed shortcut commands to invoke predefined AI assistant workflows.
dialectic-partner
Run structured dialectic debates between main and sub-agents to test assumptions.
foundational-principles
Establish disciplined reasoning and execution rules for AI coding workflows.
memory-bank
Maintain persistent project context across sessions using structured .memory-bank files.
testing-guidelines
Define test design principles for coverage, clarity, independence, and maintainability.
zellij-control
Inject keystrokes and read rendered screen output through Zellij sessions.
cline-kanban
Manage coding-agent tasks on a Cline Kanban board via the kanban CLI.
help-me-read
Converts URLs into paragraph-preserving reading drafts with optional translation and commentary.
response-guidelines
Enforce concise, intent-preserving responses with optional yes/no clarification.
sequential-thinking
Iteratively generate numbered thoughts to solve complex problems via the sequentialthinking MCP tool.