Claude Prompts for Job Descriptions That Attract Right Candidates
20 copy-paste Claude prompts for writing JDs that filter rather than describe — level-calibrated, scope-specific, inclusive, and convertible into interview rubrics. Claude's structured output strength shines here.
In short: This page contains 14 copy-paste ready prompts, organized into 4 categories with a description and pro tip for each. The first 5 prompts are free instantly, no signup needed. Hand-curated and tested by the AI Academy team.
JD Writing
4 promptsJD from Hiring Manager Brain Dump
1/14✨ What it does
Claude turns [Paste hiring manager's rough notes about the role] into a structured JD with role summary, day-to-day and quarterly work, required and preferred experience, 90-day and 1-year success, reporting, comp range, and work style. Drop in the notes, then fill the comp and reporting facts before you post.
[Paste hiring manager's rough notes about the role]. Write a structured JD. Output as XML: <role-summary>, <responsibilities-day-to-day>, <responsibilities-quarter>, <required-experience>, <preferred-experience>, <success-after-90-days>, <success-after-1-year>, <reporting-structure>, <compensation-range>, <work-style>, <growth-path>. Specific over generic.
Pro tip: XML output = easier to slice into ATS fields, internal docs, public posting. Different audiences see different sections without copy-paste pain.
90-Day + 1-Year Success Definition
2/14✨ What it does
Claude defines 90-day and 1-year success for [role] as measurable outcomes (deliverables, relationships, impact, scope), not a list of activities. Name the role, then keep only outcomes you would actually check at those dates.
Define success at 90 days + 1 year for [role]. Output: 90-day measurable outcomes (specific deliverables / relationships built / understanding gained), 1-year measurable outcomes (impact, capabilities developed, scope expansion). Outcomes-based not activities-based — what changes in the world if they succeed.
Pro tip: Activities-based ("attended 50 1:1s") vs outcomes-based ("team Net Promoter Score improved 15 pts") differs starkly. Outcomes attract operators; activities attract clock-watchers. Choose your filter.
Differentiate Required vs Preferred
3/14✨ What it does
Claude critiques [Paste my current required + preferred lists]: which required items are really preferred, which preferred items are truly required, and what is missing, aiming for a tight required list. Paste both lists, then move anything that would lose you a good candidate.
[Paste my current required + preferred lists]. Critique: which "required" items are actually preferred? Which "preferred" items are actually required if I'm honest? What am I missing? Goal: tight required list (otherwise I lose qualified candidates), generous preferred list (signal range without filtering).
Pro tip: Inflated required lists = women + underrepresented apply only if they meet 100%; men apply at 60% match. Tight required = wider candidate pool with no quality drop.
Scope-Calibrated Description
4/14✨ What it does
Claude rewrites the JD for [role] at [company stage / size] so scope is honest: budget, team size, decision authority, ambiguity, and generalist versus specialist. Fill both blanks, then use the draft to filter people who want a different stage than you.
I'm hiring [role] at [company stage / size]. The same role is different at startup vs scale-up vs enterprise. Help me write JD that signals scope correctly: budget responsibility, team size, decision authority, ambiguity tolerance, generalist-vs-specialist orientation. Filter mismatched candidates upfront.
Pro tip: Senior person from BigCo at startup often fails — wants infrastructure + specialists. Senior person from startup at BigCo often fails — wants speed + autonomy. JD that signals scope filters mismatched applicants.
XML tags are just the start. Learn the full Claude workflow.
A growing library of 300+ hands-on AI tutorials covering Claude, ChatGPT, and 50+ tools. New tutorials added every week.
Inclusive Language + Bias Audit
3 promptsInclusive Language Audit
5/14✨ What it does
Claude audits [Paste JD] for exclusionary language: gendered slang, age-coded phrases, ableist requirements, extra credentials, insider jargon, and inflated years, and suggests replacements. Run the audit, then accept the flags that would shrink your pool for no reason.
[Paste JD]. Audit for biased or exclusionary language. Flag: gendered terms (rockstar, ninja, guru), age-coded language (digital native, energetic), ableist phrasing (must stand for X hours unless essential), academic-credentialism unnecessary for role, jargon excluding outsiders, unrealistic year requirements. Suggest replacements.
Pro tip: Tools like Textio quantify this. Claude can do similar scan + suggest replacements. Cleaner JDs = wider applicant pools = better hires (and legally cleaner).
Years-of-Experience Reality Check
6/14✨ What it does
Claude pressure-tests a JD that requires [X years experience]: whether the skill really takes that long, what could substitute, and what to drop as gatekeeping. Enter the years, then prefer the junior with the right skills over a long mediocre resume you would regret.
JD requires [X years experience]. Pressure-test: is this real (skill takes that long) or copy-paste from old JDs? What can substitute (e.g., 3 years + advanced cert = 5 years)? Drop unrealistic requirements that gate-keep without justification. Junior candidates with right skills > 7-year mediocre.
Pro tip: Most "5+ years" requirements = unjustified. Specific skill mastery + scope handled > arbitrary year count. Ask "would I hire someone with 3 yrs but X demonstrated?" — if yes, drop the year requirement.
Salary Transparency Calibration
7/14✨ What it does
Claude drafts a published comp range for [role] at [company] using market, internal equity, and seek-rate, plus what to say about variable pay and equity. Fill both blanks, then publish the range you can actually hire inside.
Help me draft the comp range for [role] at [company]. Considerations: market data, internal equity, candidate seek-rate at different bands. Output: range to publish, rationale for top vs bottom of range, what variable comp + equity I should mention, language framing the range honestly.
Pro tip: Wide ranges ($90K-$180K) signal mistrust. Tight ranges signal honesty + filter mismatched candidates. Pay transparency laws are leveling the playing field anyway. Be early > be forced.
JD-to-Rubric Conversion
3 promptsJD → Interview Rubric
8/14✨ What it does
Claude converts [Paste JD] into an interview rubric: competencies, level behaviors, sample questions, passing and failing answers, and a 1-4 scale with anchors, as XML for an ATS. Start from that JD, then brief interviewers on the anchors before your first loop.
[Paste JD]. Convert into structured interview rubric: required competencies, behaviors that demonstrate each at level, sample interview questions per competency, what passing answer looks like, what failing answer looks like, scoring scale (1-4) with anchors. Exportable as XML for ATS.
Pro tip: Rubric created at JD-time = consistent hiring. Rubric improvised at interview-time = bias. Same JD writer should write rubric — alignment built-in.
Take-Home Assignment from JD
9/14✨ What it does
Claude designs a take-home from [Paste JD] that tests the real job in under 4 hours, is evaluable, and is not free labor, with a brief, criteria, and time guidance. Time yourself on the assignment, then send it only if your clock stays under four hours.
[Paste JD]. Design a take-home assignment that tests the actual job. Constraints: under 4 hours of candidate time, simulates real work (not trick puzzle), evaluatable against rubric, doesn't exploit free labor. Output assignment brief, evaluation criteria, time guidance.
Pro tip: 4 hours = ethical limit. Anything longer = exploitation. "Real work simulation" > "leetcode puzzle." Test the actual job, not unrelated CS trivia.
Reference Check Questions from JD
10/14✨ What it does
Claude writes 8-10 reference questions from [Paste JD] on the listed competencies, plus stretch and red-flag questions beyond whether they were employed. Use those questions on the call, then skip the generic form that does not test your role.
[Paste JD]. Generate reference check questions tailored to this role. Output: 8-10 specific questions on competencies that JD lists, calibration questions ("what stretch did you see them grow into?"), red flag questions ("what would have been their ideal next role you didn't have to offer them?"). Beyond verifying employment.
Pro tip: Generic reference check = "are they good?" = useless answers. Specific question on the JD's competencies = useful signal. Tailored questions = the difference between confirming a hire + catching a bad hire.
These prompts give you the what. Tutorials give you the why.
Learn when to use extended thinking, how to build Claude Projects, and workflows that compound. 300+ tutorials and growing.
Internal + External Posting
4 promptsExternal Job Posting Adaptation
11/14✨ What it does
Claude adapts [Paste internal JD] for an external posting: company context, why the role matters, a compelling opener, less jargon, clear apply steps, and an EEO line. Expand the internal doc, then add the context an outsider would not have at your company.
[Paste internal JD]. Adapt for external posting. Different from internal: more company context, more "why this role matters," compelling opener, role intrigue + transparency, less internal jargon, application instructions clear, EEO statement. Internal docs assume context external candidates don't have.
Pro tip: Internal JDs written for HR + manager. External JDs written to compete for attention. Same role, different document. Most companies confusingly post internal version externally.
LinkedIn Job Post Variant
12/14✨ What it does
Claude rewrites [Paste external JD] as a 150-250 word LinkedIn post: conversational opener, problem-led framing, 3 requirement bullets, and one CTA. Shorten the corporate page, then post the version a scroller on your feed would finish.
[Paste external JD]. Variant for LinkedIn: shorter (150-250 words), conversational opener, problem-led framing ("we have X problem; you solve it"), 3 bullets max for requirements, single clear CTA. LinkedIn scrolls past 5 pages of corporate JD.
Pro tip: LinkedIn = scroll. 5-page JD scrolled past. 250-word problem-framed post = clicked + applied. Same role, different attention pattern, different copy.
Internal Posting + Promotion Path
13/14✨ What it does
Claude writes an internal posting for [role]: how internals are evaluated versus externals, who should apply, manager support, timeline, the committee, and how you will support people who are not selected. Name the role, then share the process before people apply.
Internal posting for [role]. Existing employees can apply. Output: how internal candidates are evaluated (vs external), encouraged candidate profiles, what current managers should support, application timeline + transparency, who's on hiring committee, how non-selected internal candidates will be supported afterward.
Pro tip: Internal applicants who don't get the role = often quit if not handled well. Transparency + support afterward = retention. Treating them like outside candidates with no follow-up = surprise resignations.
JD Sunset Check
14/14✨ What it does
Claude audits [Paste JD that's been open 60+ days] for whether the posting is the problem: market fit, comp, over-built requirements, weak narrative, or company perception. Diagnose the stale listing, then fix the posting before you leave it open longer.
[Paste JD that's been open 60+ days]. Help me assess: is the JD itself the problem? Audit: realistic for talent market, comp competitive, requirements over-engineered, role narrative compelling, company perception. Sometimes the JD is broken; opening longer doesn't fix it.
Pro tip: Roles open 90+ days = JD-fix > more sourcing. Re-write often unlocks pipeline. "We can't find anyone" = often "we wrote a unicorn JD that doesn't exist."
Free tool
Resume Bullet Rewriter
Turn a weak resume bullet into 3 strong, metric-driven versions.
Frequently Asked Questions
Prompts are the starting line. Tutorials are the finish.
A growing library of 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ AI tools. New tutorials added every week.
7-day free trial. Cancel anytime.
Related guides