An AI-powered interview preparation and candidate assessment toolkit you can adapt to any role, any company, any domain. Load your project context, drop a resume, get a tailored interview plan and post-interview analysis.
Four steps from job description to hiring decision. Claude Code reads CLAUDE.md + your project context and does the rest.
Company, role, JD
and key signals
PDF, DOCX, or
pasted text
Tailored questions
in a Word document
Full analysis, scoring
& verdict generated
Drop the following into context/ so Claude understands what role you are hiring for:
Claude Code reads your project context plus the framework rules and produces professional deliverables automatically.
Tailored questions across 6 sections based on the candidate's resume and your role context. Targets gaps, validates claims, includes red flags and a scoring sheet.
Section-by-section breakdown with direct quotes. Three universal tests: Honesty, Role Readiness, Analogical Thinking. Scored on 9 calibrated criteria with a clear verdict.
Every new candidate is compared against everyone you have already interviewed for the same role. The comparison table updates automatically.
Drop an MP3 or WAV recording and Claude transcribes it with Whisper, then runs the full analysis automatically.
Interview guides can be emailed to other interviewers via Resend. Formatted HTML with the scoring sheet included.
The framework adapts to your role. Whether you are hiring a backend engineer, a sales lead, a data scientist, or a domain specialist, the questions match your context.
9 criteria scored 1–5, averaged into a final verdict. Override or extend any criterion in your project context.
| Criteria | What We Evaluate |
|---|---|
| Domain Depth | Hands-on knowledge of the role's primary domain (frameworks, tools, methodologies) |
| Technical / Functional Skills | Certifications, breadth of tooling, depth of execution, architecture or process design |
| Role-Specific Capability | The signature skill of the role (e.g. selling, coding, designing, leading, analyzing) |
| Verified Achievements | Specific numbers, decomposed claims, credible project descriptions with scale |
| Personal vs. Brand | Did this person deliver outcomes themselves, or ride a strong company brand? |
| Competitive Awareness | Understanding of alternatives in the market and how to position against them |
| Culture & Context Fit | Mindset, growth orientation, alignment with the company's stage and style |
| Honesty & Self-Awareness | Volunteers weaknesses, describes failures, corrects inaccuracies |
| Communication | Reads the room, calibrates to audience, tells stories with specifics |
Patterns that hold up whether you are hiring an engineer, a seller, a designer, or a domain expert.
Candidates who volunteer weaknesses, describe failures, and correct themselves consistently outperform polished but evasive ones. Authenticity is hard to fake under pressure.
Every role has one core skill. Ask about it directly. If the candidate pivots away from it repeatedly, they do not have it — regardless of what the resume claims.
"Walk me through your last release / deal / design / investigation" produces 10x more signal than open-ended "tell me about" prompts. It forces specifics.
Strong candidates can decompose their claims (team size, time period, personal contribution, outcome metric). Weak candidates give round numbers or no numbers at all.
The more "scalable, end-to-end, stakeholder alignment" appears in every answer, the less likely there is real depth. Follow up with: "Give me the actual number."
A candidate with the right mindset can learn a new domain in weeks. A candidate with poor self-awareness or communication will not improve regardless of training.
Everything is organized for Claude Code to navigate automatically.