baijidot/bytedance-offer-copilot-skill

ByteDance Offer Copilot is a Python tool aimed at Chinese university students preparing for job interviews, particularly at technology companies like ByteDance. The tool is designed to give honest, critical feedback rather than flattery. Its central framing is that it roleplays as a senior ByteDance interviewer who identifies vague language, empty claims, and weak project descriptions in resumes and interview answers. The tool has 14 modules covering different parts of the job application process. You can paste in a job description and get a breakdown of what the company is actually looking for and a 30-day preparation plan. You can submit a resume and get an offer probability score across seven dimensions with specific gaps identified. There is a resume rewriting module that converts student-sounding descriptions into phrasing more typical of tech industry professionals. A separate module detects what the README calls "hollow corporate buzzwords": vague verbs, unquantified claims, and empty phrases, then flags them with explanations. The mock interview module supports four modes ranging from gentle to high-pressure. In the high-pressure mode, a dynamic stress score is calculated based on how you answer: vague expressions increase the score, short responses raise it more steeply, while honest admissions of uncertainty and evidence of structured thinking lower it. There is also a contradiction detector that reviews your answers across multiple rounds and flags inconsistencies, and a group interview simulator where the AI plays several different participants in a leaderless discussion scenario. Additionally, the tool can analyze whether a project described on a resume looks like a real product with actual users or a demo created for coursework, and it can match your background to specific role categories with concrete gap analysis. The architecture uses a two-layer approach: simple rule-based checks for instant feedback (buzzword detection, authenticity signals) and language model calls for deeper analysis like interview follow-up questions and evaluation reports. The project is released under an MIT license and requires Python 3.9 or later.