Résumé · Toolkit
The résumé-tailoring agent
Every job-specific résumé on the previous page was generated by this system against a real posting. Rather than claim that, here is the system prompt, the process it runs, and what it returns — including the part most tools hide, which is what the résumé cannot support.
The system prompt
You tailor one master résumé to one specific job posting. INPUTS master_resume — the canonical résumé, in full posting — the job posting, verbatim constraints — page limit, format, any client instruction RULES 1. Never invent. Every claim in the output must appear in master_resume. You may re-word, re-order, re-scope and cut. You may not add. 2. Rank by evidence, not by keyword. If the posting asks for something the master does not evidence, say so in the gap list rather than implying it. 3. Lead each bullet with the outcome, then the mechanism. A bullet that describes a responsibility with no result is a bullet to cut. 4. Mirror the posting's vocabulary only where the master already supports the concept. Matching a keyword you cannot back is how a résumé fails an interview rather than a filter. 5. Keep it to the page limit by deletion, never by shrinking type. OUTPUT resume — the tailored résumé rationale — what was promoted, demoted or cut, and why gaps — requirements in the posting the master does not evidence
The process
01
Parse the posting
Requirements are separated into must-have, nice-to-have and noise, and each is tagged with the concept behind it rather than the phrase used. "Marketing automation" and "lifecycle tooling" are one requirement, not two.
02
Score the master against it
Every bullet in the master résumé is scored for evidence against each requirement. This is the step that decides the document — ordering falls out of the scores rather than out of taste.
03
Draft under the page limit
The highest-scoring evidence is promoted and the rest is cut. The constraint is deliberate: a one-page limit forces the ranking to mean something.
04
Report the gaps
What the posting wants and the master cannot evidence is returned as a list. That list is the useful output — it is either what to address in the cover letter, or the reason not to apply.
Worked drafts
Each of these was produced by the process above against the posting linked beside it — so the output can be read against the brief it was written for.
Why the gap list matters
A tailoring tool that never returns a gap is telling you it is willing to imply things the résumé cannot support. The gap list is the honesty constraint made visible — and in practice it is the output I use most, because it decides whether the application is worth making at all.