About

Making technical hiring more honest

Resume inflation is a well-known problem in tech. skill·proofer gives hiring teams a fast, automated way to separate claims from evidence before the interview.

Why this exists

Technical interviews are expensive. A single bad hire costs months of productivity. Yet most screening relies on self-reported resumes and generic interview questions that don't verify actual experience.

We built skill·proofer to give hiring teams a third-party verification layer — grounded in real code, not self-promotion. The output is a report that tells you what a candidate has actually built, how they work, and what questions to ask in the interview.

It's not a replacement for interviews. It's a tool to make them more effective.

How it works

The pipeline has 5 steps. Two use AI; the rest are deterministic — same input always produces same output.

01

CV Parser

AI

Extracts technical claims from the uploaded PDF using AI. The output is a structured list of skills, technologies, and experience claims to verify.

02

GitHub Fetch

Fetches public repositories, commit history, language breakdowns, and README content via the GitHub API. No authentication required — all data is public.

03

Evidence Match

Matches each CV claim against repository content. Checks READMEs for keywords, language byte counts for proficiency claims, and commit patterns for experience depth.

04

Fingerprint

Computes a deterministic technical profile: primary language, activity level, solo vs. collaborative work style, commit frequency, and recency.

05

Question Gen

AI

Generates 5 interview questions per important repository — 2 general (architecture, design choices) and 3 code-specific (referencing actual files and snippets). Each question includes what a strong answer would mention.