What problem does it solve? Preparing for technical interviews requires turning real project work into resume-worthy highlights, anticipated interviewer questions, and rehearsed answers, which is time-consuming and often produces vague, unverifiable claims. ## Core Features & Use Cases - Evidence-Based Highlight Mining: Scans README, dependencies, git history, and source code to extract verifiable project highlights across architecture, technical challenges, performance optimization, design patterns, and engineering practices, rejecting anything without code evidence. - Interviewer-Perspective Question Generation: Produces structured questions per highlight across five types (follow-up, boundary, trap, extension, fundamentals), with coverage checks requiring at least one follow-up, boundary, and trap question per highlight. - Project-Grounded Reference Answers: Generates recitable answers tied to concrete implementation details, tagged by depth ([核心]/[进阶]/[加分]) with follow-up contingency plans and honesty about limitations. - Use Case: A backend engineer preparing for senior-level interviews points the Skill at their repository and receives three cross-referenced Chinese Markdown documents (highlights.md, questions.md, answers.md) in an interview-prep/ folder, linked by shared highlight IDs. ## Quick Start Ask the agent to analyze this repository and produce resume highlights, interview questions, and reference answers as three Markdown documents in an interview-prep folder.
