Recruitment Screening Prompt Pack
A set of production prompts used by the AI hiring agent. Copy, adapt the bracketed parts, and run them in order.
1. Extract a rubric from a job description
code
You are a hiring analyst. Convert the job description below into a scoring rubric.
Output JSON: { "must_haves": [...], "nice_to_haves": [...], "disqualifiers": [...],
"weights": { requirement: 1-5 } }. Exclude anything not job-relevant
(age, gender, nationality, university prestige).
JOB DESCRIPTION:
[paste JD]
Build in Claude
Build in ChatGPT2. Score a candidate with evidence
code
Score this candidate 0-100 against the rubric. For EVERY requirement, quote the exact
line from the CV that supports your judgement, or write "no evidence". A score without
quoted evidence is invalid. Output: { score, per_requirement: [...], summary }.
RUBRIC: [paste rubric JSON]
CV: [paste candidate CV text]
3. Generate interview questions for a shortlisted candidate
code
Based on this candidate's gaps versus the rubric, write 5 interview questions that
probe the weakest areas. Make them behavioural and specific to their experience.
4. Bias self-check
code
Review your shortlist. List any criterion you applied that is not in the approved
rubric. Flag any pattern that could create adverse impact on a protected group.
Usage notes
- Always run prompt 4 before sending a shortlist to a hiring manager.
- Keep raw outputs - they are your audit trail.