mikezhangyl/jd-cv-grill

JD CV Grill is a bilingual skill for AI coding assistants that helps a person tailor their CV or resume to a specific job description. It is meant for experienced candidates who already have plenty of material, but who freeze up when they see a new job posting and cannot tell which of their past projects matter for this role. The skill works in both English and Chinese and supports any mix of the two between the candidate, the resume, and the job description. The author is direct about what it is not. It is not a tool for inventing experience or padding a resume with achievements that did not happen. Instead it pushes the AI agent to read the job description carefully, pull out the hard requirements, the hidden hiring signals, and the keywords that automated tracking systems look for, then interview the candidate about real past work. The interview style is one question at a time, not a long form to fill in. As the candidate answers, the agent builds an evidence table that lines up each job requirement against a piece of personal proof, notes how strong that proof is, and flags missing details, confidentiality issues, or numbers the candidate cannot verify. Resume bullets are then rewritten around confirmed facts, with attention to personal contribution, scope, impact, tools used, and constraints faced. Installation is handled by the Agent Skills CLI. The README shows npx skills add mikezhangyl/jd-cv-grill commands for Codex, Claude Code, Cursor, or several agents at once, with a -g flag for a global user install and a --copy flag to copy files instead of symlinking. There is also a manual git clone path for cases where the CLI is not available. Once installed, the user invokes it inside their AI agent with a trigger like $jd-cv-grill plus a request to tailor a resume to a job description. The repository itself is small: a SKILL.md file plus an agents/openai.yaml configuration, MIT licensed.