- Home
- Skills
- HR & Recruiting
- Interview Question Bank Builder
Works with the AI Agents you already use
Interview Question Bank Builder
The output is two files from one source: a markdown bank an interviewer can print and a JSON bank a tool can read.
Free
Interview Question Bank Builder
Example session with this skill installed
Build a 12 question behavioral interview bank for a senior product designer from this job description, with anchored rubrics and a scorecard.
- Read your context and instructions
- Compiled the interview question bank
Senior Product Designer — Behavioral Interview Bank
How to use this bank
Questions are grouped by competency. Each competency is traced to a quoted line from the job description. Score each question on the 1 to 5 anchored rubric, then total the candidate scorecard at the end. Timing fits a 60-minute interview.
Competency: Product judgment
JD line: [quoted job-description line for this competency]
Question 1
Tell me about a product decision where you changed direction after learning something new about users. What did you change, and what happened next?
Rubric
- 1: Describes a decision with no user evidence and no outcome.
- 2: Mentions user input but does not connect it to the decision.
- 3: Explains a clear decision, the evidence behind it, and the result.
- 4: Explains trade-offs, alternatives considered, and a measured result.
- 5: Explains trade-offs, evidence, stakeholder alignment, and a lasting product or user outcome.
Answer sketch: A strong answer names the original direction, the new evidence, the revised direction, and the result.
Red flags: No evidence, no outcome, or credit taken for another person's decision.
Follow-up probe: What would you do differently with the same evidence today?
Candidate scorecard
Record a 1 to 5 score for each question, grouped by competency, with notes tied to the rubric anchors. Do not use this kit to make a hiring decision on its own: the skill builds the kit and the humans make the call.
Bias scan: Passed. Quality gate: Passed. Output files: markdown bank for printing and JSON bank for tools, rendered from the same validated bank.
Connects securely to your tools. The creator never sees your data.
About this skill
Turn a job description into an interview kit: competencies traced to the exact lines of the JD, questions grouped by competency, and a 1 to 5 anchored rubric for every question, so two interviewers score the same candidate the same way.
For interview questions, hiring rubrics, candidate scorecards, interview guides, and help running a hiring loop for a role, at any seniority from intern to executive, for phone screens, technical rounds, behavioral interviews, and final culture rounds, including final rounds with company values. The output is two files from one source: a markdown bank an interviewer can print and a JSON bank a tool can read.
Why it matters: Without anchored rubrics, two interviewers score the same candidate differently and the loop becomes opinion. With competencies grounded in the JD, a question for every competency, a rubric anchor for every question, an answer sketch, red flags, a follow up probe, and a scorecard, the loop stays consistent and defensible.
What it does:
- Validates the job description, seniority, round, and count, and warns when the JD's own level words contradict the seniority picker.
- Extracts competencies only where they can be tied to a quoted JD line. A competency that cannot be tied is cut or moved to the assumptions log.
- Builds questions grouped by competency, each with an anchored 1 to 5 rubric, answer sketch, red flags, follow up probe, and candidate scorecard.
- Runs a bias scan over question text, intent note, follow up probe, answer sketch, every red flag, every rubric anchor, and JD quotes. A bank with a surviving hit in a scanned field is a failed build, not a warning.
- Runs a quality gate checking bank shape, competency grounding, rubric anchor quality, duplicate questions, timing arithmetic, and rerunning the bias guard.
- Renders the markdown bank and JSON bank from one validated bank, and refuses to write anything if validation fails.
- Keeps timing inside the interview slot: default 30 minutes for a screen round and 60 otherwise, settable to 30, 45, or 60.
Honest limits:
- It writes interview kits. It does not make hiring decisions. It never scores a real candidate, recommends a hire, or writes a rejection. If asked, it says the skill builds the kit and the humans make the call.
- Not for writing candidate rejection emails, reference checks, compensation decisions, or hiring decisions themselves.
- The bias guard is a floor, not a substitute for judgment. It catches questions about protected characteristics, protected life circumstances, salary history, genetic and family medical information, statutory leave, and prior discrimination complaints, plus seven lawful but bias prone framings. It cannot catch paraphrase. Use judgment on top of the scan.
- The scan is pattern matching over a fixed phrase list, not legal review. It is a floor: necessary, not sufficient.
- The compliance reference is guidance for US hiring and it changes. State and local law varies. This is not legal advice.
- A JD under 25 words is not enough to work with. Around 150 words or more is comfortable.
- No network, no API keys, no setup. Python 3 standard library only.
Requirements: A job description, seniority, and round type. Optional question count, default 12, range 6 to 24, interview minutes, and company values for the final round.
Example prompt: Build a 12 question behavioral interview bank for a senior product designer from this job description, with anchored rubrics and a scorecard.
What is in the package: SKILL.md, LICENSE, references including bias and compliance, bank schema, examples, scripts for input validation, bias scanning, bank validation, and rendering, and tests.
It writes interview kits only; it does not make hiring decisions, score real candidates, or write rejections. The bias scan is pattern matching, not legal review, and cannot catch paraphrase. The compliance reference is guidance for US hiring only; state and local law varies. A JD under 25 words is not enough; around 150 words or more is comfortable.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 23 days ago
- Free with an Agensi account
Requires