Resume Screening & Ranking Assistant

The problem this kills

A single posting can pull in 200 resumes. Skimming them by hand is slow, and worse, it's inconsistent: the tenth resume gets a fairer read than the hundredth, and the criteria quietly drift as you get tired. You end up with gut-feel decisions you can't fully explain later - which is exactly the kind of thing that gets a hiring process in trouble.

Generic AI resume tools swing too far the other way: they auto-reject candidates behind a black box, and you have no idea why someone got filtered out. That's both legally risky and a great way to throw away good people.

This plan builds the middle path: AI does the tireless first pass - reading, summarizing, and scoring every resume against your job-related criteria - and then hands a ranked, evidence-backed shortlist to a human who makes the actual call.

What you'll build

A private internal web app for your recruiting team that:

• Lets you define a scorecard for a role: must-have skills, minimum years, required certifications, location / work-authorization needs, and weighted nice-to-haves.
• Accepts resumes in bulk - PDF/DOCX uploads or pasted text - and parses each into a structured candidate summary.
• Produces a transparent match score for each candidate, with the specific lines of evidence the score is based on and a confidence flag when the AI isn't sure.
• Ranks candidates into a reviewable list, where the recruiter sees the score and evidence side-by-side with the resume and can accept, reject, or edit every result.
• Saves the shortlist with a full audit trail and exports a clean CSV with scores and reasons you can hand to a hiring manager or paste into your ATS.

What's inside the Implementation Plan

The plan is a complete, paste-and-go runbook for Claude Code. It opens by interviewing you about your actual recruiting process - your roles, your ATS, how your resumes really look, your volumes, and your decision rules - and reads back a short tailored spec for your thumbs-up before it builds anything. So you get a tool shaped around how you actually hire, not a generic template.

From there it walks you, prompt by prompt, through standing up the database, login, the scorecard builder, the bulk-upload and parsing step, the AI scoring with cited evidence, the recruiter review screen, the audit trail, and the CSV export. Every build step ends with a ready-to-copy prompt. There's a "No API yet?" fallback throughout, so you can build and use the whole thing today by pasting resume text and exporting a scored CSV - no integration required.

The governance it includes (this is the point)

• Login so only your team can open the tool.
• Row-level security so a recruiter only ever sees their own organization's roles and candidates.
• A human-in-the-loop gate: the AI ranks and explains, but a person makes every accept/reject decision. The AI never auto-rejects.
• A complete audit trail: who scored, who decided, what they changed, and when.
• Duplicate guards: the same applicant submitted twice (by email, or name + phone) is merged, not double-counted.
• Bias guardrails built in: scoring is restricted to job-related criteria, protected characteristics are never used, and every score shows its reasons so decisions are defensible.

Who it's for

Recruiters and hiring managers who today skim resumes by hand and want a faster, more consistent first pass without handing the decision to a machine. If you can fill out a job scorecard and read a resume, you can run this tool - and with this plan, you can build it.

You've got this - paste the first prompt and let the agent interview you.