An Open-Source AI Skill That Rewrites Developer Resumes with Interviewer-Level Rigor
Resume screening in China's tech hiring market is brutal and formulaic. This skill operationalizes the unwritten rules that separate a rejected CV from one that lands an interview — rules that apply just as sharply to Western developers competing in a market where AI-assisted screening is now the norm.
The skill, packaged as a plain-text prompt installable into any standard AI agent, reads a developer's PDF resume and produces a markdown rewrite that follows strict rules: replace vague self-assessments with specific, quantified achievements; expand three-line skill sections into six dimensions; and restructure project entries into a "verb + technology + result" format. It also strips out filler like age, years of experience, and generic personality summaries that HR already infers or ignores.
A companion console-based agent called Zhiyu handles the full pipeline: it reads the local PDF, applies the skill, outputs a markdown file, and can generate a styled PDF automatically. The skill is agent-agnostic and works with OpenClaw, Claude-code, Cursor, or any tool that loads skills from a standard `.agents/skills` directory.
Future iterations plan to bundle example resumes, PDF templates, and a mock-interview skill covering common Chinese tech-interview beats: self-introduction, reasons for leaving, strengths and weaknesses, employment gaps, and project deep-dives.
Encoding domain expertise as a reusable skill prompt is a lightweight alternative to fine-tuning or building a dedicated SaaS tool; the entire "product" is a text file that slots into any agent runtime.
The skill's insistence on deleting age and years of experience reflects a specific Chinese hiring-practice nuance — HR infers seniority from graduation dates, and stating it explicitly can work against a candidate — that may surprise Western developers accustomed to including both.
Structuring project bullet points as "verb + technology + result" is not novel advice, but baking it into an automated rewrite pipeline turns a manual editing chore into a one-shot generation step, which changes how a developer approaches the first draft.