Frequently asked questions

DevStamp analyzes your real GitHub repositories and turns your experience into an honest, evidence-based rating for the skills you choose, proof you can share with recruiters instead of another self-reported line on a resume.

Three steps: sign in with GitHub, pick a repository and the skills you want to verify, and DevStamp reads the relevant code and its commit history to produce a rating for each skill, backed by a written justification.

DevStamp focuses on the code that actually matters for each skill rather than reading every file blindly. For each skill it selects the files most relevant to it and analyzes them in depth, together with your repository's full commit history: the timeline, commit messages, and how the relevant code evolved over time, across the whole project.

To keep each analysis focused, there's a limit on how much file content is read in a single pass, so on large repositories we read the most relevant files rather than every file, but the commit-history signals always span the entire repository. On the results page you can see exactly which files were analyzed for each skill.

When a skill scores highly with little sign of unreviewed AI, a deeper audit reads additional files and specific commits to confirm the result.

Each skill is rated on the depth of demonstrated experience:

• 0 (Not found): no code related to the skill was found in the repository at all.
• 1 (No evidence): the skill isn't meaningfully used.
• 2 (Familiar): basic or tutorial-level usage.
• 3 (Proficient): real, working usage with understanding.
• 4 (Advanced): production-grade work with clear ownership.
• 5 (Expert): deep, architecture-level command of the skill.

No. Using AI assistants is normal in modern development and we don't penalize it. What we flag is AI output that was shipped unreviewed: code that was never read, corrected, or made to fit the project, showing up as incoherence, broken wiring, or leftover AI boilerplate. Well-directed AI whose output is coherent and iterated on is fully credited; a rating is only reduced when the work shows the AI ran unchecked.

DevStamp only accesses the repositories you choose to analyze, through GitHub's standard authorization. Your code is cloned temporarily to run the analysis and then deleted, it isn't stored or shared. Private repositories are only included if you explicitly grant access.

A certificate is a shareable, public page that showcases a set of your verified skills and their ratings. To create one, go to My Certificates → Create Certificate, give it a title (e.g. "Backend Developer"), and pick which of your completed verifications to include. You'll get a public link anyone can open, no login required to view it.

DevStamp certificates fit LinkedIn's "Licenses & certifications" section. Open your certificate page first and keep its link and ID handy, then:

1. Go to your LinkedIn profile and click Add profile section.
2. Under Recommended, choose Add licenses & certifications.
3. Name: enter your certificate title (e.g. "Backend Developer").
4. Issuing organization: type DevStamp.
5. Issue date: the date shown on your certificate.
6. Credential ID: the Certificate ID shown at the bottom of the certificate page.
7. Credential URL: paste your public certificate link (the /verify-certificate/… page). This makes it clickable so anyone viewing your profile can verify it.
8. Click Save. It now appears in your Licenses & certifications section.

A wide range: programming languages, frameworks, databases, cloud and DevOps tools, and more. When you start a verification you can search the built-in list, and request any skill that isn't listed yet.

Yes. If a skill in a session ends up as Failed, the results page shows a Retry Failed button that re-runs just the failed skills.

Each skill analysis costs €2, or €4 with Deep Verification. You can also request a free analysis from the verification page, we review each request and, if approved, email you a link to your result.