Resume Tailor: Delegate Chinese Resume Formatting to Agents

Where this fits

You have a Chinese resume in an old Word file. You want the clean ASu single-column HTML/PDF layout, but you do not want an agent rewriting the internship bullet points you spent weeks tuning. resume-tailor solves this with a forced workflow: ask first, preserve confirmed facts, and only restructure into a fixed template. If you are starting from zero, it builds experience and project sections from experience-deepthink interview transcripts instead of inventing filler.

Why agents benefit

• It starts by reading installation-mode and voice-contract files, so it knows whether Feishu-only delivery applies and can adjust expression without touching facts.
• A mandatory first question—"Do you have an existing resume?"—routes into preservation mode or from-scratch mode, preventing mixed workflows.
• From-scratch mode requires user-selected modules and order, and treats user-supplied materials as candidate project experience, not as confirmed content.
• It converts to Markdown first and asks for confirmation before generating editable HTML and a max two-page A4 PDF, giving a safe checkpoint.
• The ASu template is read-only; the agent copies it per user, so output stays consistent without redesign.

Practical scenarios

Migrating an existing resume without content drift

Attach a current resume; the agent copies the read-only template to a user-specific HTML file, moves facts into the single-column layout, and does not improve wording or remove experiences. Output is a faithful reformat.

Building a resume from interview transcripts

During interview prep you have long experience-deepthink verbatim notes. The agent uses those as the source for project descriptions, asks which modules to include—education, experience, projects, self_evaluation—and confirms order before generating files.

Updating for a new target role

Supply a JD and choose experiences explicitly. The agent uses the voice contract to make phrasing consistent with your habits, keeps selected experiences, avoids choosing for you, and delivers Markdown, HTML, and PDF.

Add it to your agent workflow

Install it as a skill:

npx skills add riwonswain-ovo/OfferLoop --skill resume-tailor

Then a minimal turn:

User: Please convert my attached resume to the ASu single-column HTML/PDF.
Agent: Do you already have a resume whose content should be kept, only reformatted?
User: Yes.
Agent: I'll preserve all facts. First, review this Markdown version.

Compare before adopting

Compare it with other resume skills on four axes:

• Quality signals: Does the output match the ASu template? Does it enforce max two pages? Does it preserve source wording?
• Maintenance freshness: The upstream is Hisn00w/ASu-skills; check whether OfferLoop tracks it, and whether the 13-star repo has recent commits and responsive issue handling.
• Alternatives: Generic Markdown-to-PDF generators can be more flexible but lack Chinese resume conventions and experience-deepthink sourcing; full agent suites may add unnecessary user profiling gating.
• Workflow fit: The skill supports Feishu full mode for source contracts; local HTML/PDF is final delivery but records are written to the user's private knowledge base. Ensure your agent can honor that.

Why it is worth tracking

resume-tailor is worth tracking because it demonstrates a safe pattern for agent-driven document production: read runtime contracts, ask before acting, confirm intermediate Markdown, and never rewrite user-confirmed facts. Those constraints matter more than the template itself when your agent handles resumes, cover letters, or similar personal artifacts. The project is early—13 stars and one skill in the OfferLoop repo—so evaluate it when you need a Chinese resume workflow rather than replacing an existing pipeline. You can also inspect its metadata and related skills on OpenAgentSkill at https://www.openagentskill.com/skills/riwonswain-ovo-resume-tailor.