Six prompts that do the boring 80% of your job search (copy-paste)

Every job-search advice article argues about wording. Almost none of them argue about the actual bottleneck: the volume of reading, scoring, tailoring and following-up that sits between a job board and an interview. That part is language work, and language work is exactly what these six prompts are for. They are deliberately boring. That is the point: your judgment stays on the five decisions that matter, the machine carries the two hundred boring ones.

One rule before you start. These prompts are for the pipeline, not for the performance. Everything that leaves your name stays in your voice, edited by you. An interview measures you; a draft is just logistics.

1. The job description deconstructor

Read the following job description. Split it into: (a) the requirements they state, (b) the requirements they repeat or front-load, which are the real ones, (c) what this role is secretly for, inferred from team context, reporting line and the problems they describe, (d) the seniority this actually is, regardless of title. Be blunt about uncertainty. JD: [paste]

Most of the value here is (b). A description that lists "owns priorities" three times wants a different candidate than one that lists it never.

2. The honest fit score

Here is my CV and a job description. Score fit 0-100 against the REAL requirements you identified, and output: the top three evidenced matches with the CV line that proves each, the top two gaps, and a one-paragraph verdict on whether I should spend an application on this at all. Do not be encouraging. CV: [paste] JD: [paste]

The instruction that changes everything is "whether I should spend an application on this". Applications are not free; they cost the twelve minutes of tailoring and the emotional toll of the silence after. Score the role, not just the match.

3. The tailored one-page pass

Tailor my CV for this role. You may reorder, reweight and rephrase; you may NOT add experience, numbers, tools or employers that are not in my input. For every bullet you change, note which JD line it now answers. Keep total length to one page. CV: [paste] JD: [paste]

The "note which JD line it answers" trick does double duty: it gives you the tailoring and it gives you the interview prep material in one pass, because "why is this bullet here" is the second question every interviewer asks.

4. The three-sentence reach-out

Write a 3-sentence message from me to the hiring manager for this role, sent when the posting is fresh. Sentence one: the specific thing in the JD or their company's work that shows I did two minutes of reading. Sentence two: the single strongest evidenced match. Sentence three: a zero-pressure line about availability. No flattery, no "I hope this finds you well". Context: [JD + your one-line pitch]

Three sentences because hiring managers read these on a phone between meetings, and anything longer is instantly pattern-matched to mass mail.

5. The five-day nudge

Write a 40-word follow-up to my earlier application for [role]. It must add one new piece of information (a relevant result, a shipped thing, an answer to a question the JD raised) so it does not read as begging. Refer to my earlier message without apology.

Follow-ups with new information get answered; "just floating this" follow-ups teach people your name is noise.

6. The rejection post-mortem

I applied to [company/role] and got the standard rejection after [how long]. Write the two-sentence email that politely asks for one piece of specific feedback, targeted at the recruiter rather than the system, and phrased so that answering takes under a minute. Then list five patterns from my recent rejections (attached list) worth tracking.

You will get maybe one reply in ten. One reply is one datapoint, and five datapoints beat any career coach who has not seen your actual numbers. Keep the replies; they are the only ground truth in the whole pipeline.

The operating cadence that makes these work

Prompts alone just speed up bad habits. The boring 80% only pays when three operational rules ride along: apply within hours of a posting rather than days after polishing (a fresh average beats a stale perfect entry into a pile that grows by the hour), batch the pipeline reads into fixed windows instead of doom-scrolling, and keep one plain table of status for every live application so nothing rots untracked.

hunta is what it looks like when that cadence is the product instead of the habit: mailbox ingestion, scoring against the JD, tailored drafts behind a human approval gate, follow-ups on schedule, caps and cooldowns so the machine cannot send you into burnout. The demo is free and self-hosted. But the six prompts above run the honest version of this pipeline on nothing but a free LLM tier, and that is genuinely enough to change a month of searching. Your move on the other side of every draft still matters more than the tool. That part is not automatable and I would not trust anyone selling it.