The Best Resume Skill We've Reviewed Starts With Research, Not Rewriting

We imported three English-language resume skills today. They scored 9.1, 8.4 and 8.0 — a wider spread than their one-line descriptions suggest. This is the first of three posts; the others cover an ATS auditor and a seven-reviewer hiring panel.

Most resume skills are rewriting engines. You hand over a resume and a job description, and you get back the same bullets with stronger verbs. resume-tailoring does something different: it refuses to rewrite until it has researched the company and interviewed you about the experience your resume left out.

That ordering is why it scores 9.1, the highest in our directory's small job-search corner and above the 8.36 directory average.

The input is a library, not a document

The skill requires two things: a job description and an existing resume library in markdown, by default in resumes/. From there the workflow is explicit:

1. Build a searchable library from the existing resumes.
2. Research the company and role.
3. Create a template, with a user checkpoint before generation.
4. Optionally run a branching interview to surface undocumented experience.
5. Match content with confidence scores.
6. Generate MD, DOCX, PDF and an interview-prep report.

The discovery phase is the unusual part. The skill's stated mission is that "a person's ability to get a job should be based on their experiences and capabilities, not on their resume writing skills." The branching questions exist to recover the gap between what someone did and what they remembered to write down. That is a different job from editing prose, and it is the job most resume tools skip.

Why workflow scores 9.5

The evaluator called the workflow "exceptionally detailed with checkpoints, edge cases, and verification steps." The clearest evidence is the multi-job mode. When the skill detects more than one job description, it offers a single shared discovery interview instead of repeating the interview per application. It then estimates the tradeoff honestly: about 40 minutes for three jobs versus 45 sequential, and about 55 minutes for five versus 75. An 11% saving on three jobs is not a headline; the point is that the number is stated rather than assumed.

The workflow also separates research from matching from generation. Each phase has an output, and the user approves the template before the expensive step runs. That is the shape we keep rewarding in the six-dimension rubric: decisions first, prose last.

The boundary is in the description

The description states what the skill is for and, just as importantly, what it is not for:

• Not for generic resume writing from scratch (it needs an existing library).
• Not for cover letters.
• Not for LinkedIn profile optimization.

That negative space is why trigger quality lands at 8.5 rather than lower. Across the directory, skills that state a "when not to use" boundary route more reliably than skills that only list positive triggers — a pattern we have documented before.

The packaging catch

One engineering caveat is worth spelling out, because it is the kind of failure that stays invisible until an agent tries to follow the instructions.

SKILL.md references four supporting files: research-prompts.md, matching-strategies.md, branching-questions.md, and multi-job-workflow.md. In the upstream repository those files sit at the repository root, outside skills/resume-tailoring/. Installing only the skill directory — which is the normal way a skill gets packaged — produces a folder with SKILL.md and nothing else. The agent then reads instructions that point at files it does not have.

We repackaged the site download (version 1.0.1) to include all four reference files alongside SKILL.md. The evaluation still scores engineering at 8.5 because the single file depends on material it does not define. The fix is simple; noticing the dependency is the hard part.

Where it fits

The site's existing job-search skills answer narrower questions:

• review-resume — 8.1 — a product-management checklist applied to one resume in one pass.
• job-description — 8.4 — writes the job posting, not the resume.
• resume-tailoring — 9.1 — a multi-phase pipeline from a resume library to a tailored application package.

If you already have a resume and a specific role, and you are willing to spend fifteen minutes answering questions, this is the strongest option we have reviewed. If you want a five-minute opinion instead, the seven-reviewer panel is cheaper to run — and harsher.

⚡ resume-tailoring 9.1 Use when creating tailored resumes for job applications - researches company/role, creates optimi… ⬇ 1 ⚡ review-resume 8.1 Comprehensive PM resume review and tailoring against 10 best practices including XYZ+S formula, k… ⬇ 7 ⚡ job-description 8.4 Turn a hiring need into a complete job description with responsibilities, requirements, nice-to-h… ⬇ 1

The takeaway

A resume skill's first job is not sentence polish. It is deciding what evidence belongs on the page. Research and discovery do that; rewriting alone cannot. A skill that sequences the work correctly will beat a skill that merely writes better bullets, and the 9.1 reflects the sequence, not the prose.