Resume Tailor — No-Fabrication Resume Customizer
What it does
Give it your resume (txt / docx / text-layer PDF) and a job posting. It lists the posting's requirements, shows which resume lines support each one, then produces a tailored resume that only reorders, shortens or re-phrases facts you already wrote. Anything you did not write cannot appear: numbers, skills, employers, titles, dates and credentials are verified in code, and the final DOCX/PDF are re-checked against your original.
You get
…docx(editable ATS version) and…pdf(Chinese/Japanese fonts embedded, renders anywhere) — single column, standard headings, no tables, no images…_checklist.md— requirements with no evidence in your resume; every re-worded line next to its original text; lines left out; things to complete yourself…_requirements_mapping.csv,…_verification.json,…_plan.json
Languages
Traditional Chinese (繁體中文), English, Japanese — including 履歴書 and 職務経歴書 layouts. Output language = your resume's language (no translation, because translated wording cannot be verified).
Sample input → output (fake data)
Input resume line: “Cut report generation time from 45 minutes to 9 minutes by rewriting slow SQL queries.” + posting “Optimize slow SQL queries and reduce reporting latency.” → kept in the Experience section with its wording unchanged; checklist: “R2 covered ← L012”. Posting asks for “Rust or Go”, resume never mentions either → checklist: “NO EVIDENCE: R8 Knowledge of Rust or Go”, nothing added to the resume.
FAQ
- Will it add skills I don't have? No. A plan containing a word, number, tool, employer or title not in your resume is rejected and no file is written.
- Does it score me or the job fit? No. It reports evidence and gaps only.
- Does it send my resume anywhere? The Skill makes no network calls. Your Agent Client's AI provider sees whatever text your client sends it.
- Scanned PDF? Needs a text layer; there is no OCR.
- Japanese 履歴書 with photo box? It produces a text layout (name, 学歴・職歴, 免許・資格, …). Date and photo are left for you.
- Does it guarantee interviews? No.
Output format (required by Capafy Review Guidelines 2.5)
Files saved to the output folder you choose (DOCX, PDF, Markdown, CSV, JSON) and a plain-text summary in chat.
Requirements
Python 3.10+, pip install -r requirements.txt (python-docx, reportlab, pypdf, snowballstemmer, fonttools). Any Agent Client that supports Agent Skills and can run Python (e.g. Claude Code, Codex, OpenClaw, Hermes).
