← all playbooks

Brief yourself on a candidate's public work before the first call

One page on one candidate: dated work history, a line on what each recent employer does, public work where it exists, and the three questions the profile does not answer. Every fact carries its source, and the page covers the professional footprint only.

use when
you have a first call booked with a candidate and want more than a skim of their profile, without straying into anything personal
starts from
A candidate

The prompt

Paste it into Claude Code or the Claude desktop app with Hyreflow connected. The first line tells your agent to use Hyreflow, so it reads the play, asks before it spends anything, and hands the work back to you.

paste this into Claude
Use hyreflow and write me a one-page brief on a candidate before my first
call with them.

The candidate is <CANDIDATE_LINKEDIN_URL>. I am calling about
<ROLE_AND_TYPE_OF_CLIENT>.

Cover, in this order:
- Dated work history, with tenure in each role.
- One line on what each of the last <N> employers does and how big it is.
- Public work, where it exists: GitHub, a personal site, talks, articles,
  recent public LinkedIn posts about their field. If a match is not
  certain, say so and keep it out of the facts.
- The three questions the profile does not answer, each tied to a fact.

Rules. Professional footprint only. Nothing on age, health, family,
religion, politics, ethnicity or any other protected characteristic.
Nothing from personal social accounts, even if a profile links to them.
Put the source next to every fact, and write "not found" for a gap.

Tell me what the brief costs before you run the paid lookups. If I give
you several candidates, run one first and tell me what the rest cost
before you do them. Do not contact the candidate.

Replace every <PLACEHOLDER> with your own detail. Everything else can stay as written.

What you need first

  • The candidate's LinkedIn profile URL, so the brief is about the right person
  • The role you are calling about, so the questions point at it
  • A Hyreflow workspace with credits

Tools it can reach for

The agent picks per step from what your workspace has. Nothing here is required by name.

What happens when you run it

Free steps are marked free. Anything that spends credits is marked, and the agent asks before the first paid run of any size.

  1. 1

    Confirm who the brief is about

    free

    The agent repeats back the name, the profile URL and the role you are calling about, and asks for anything that is missing. Every later lookup hangs off that identity, and the handle for the post feed is read from the URL, so it is settled before anything is spent.

  2. 2

    Pull the dated work history

    credits

    The profile goes through a fixed chain of providers that stops at the first one to return roles with start and end dates. The answer carries career history and no contact details, so nothing about how to reach the candidate is bought here.

  3. 3

    Say what each employer does

    credits

    For the last few employers, one line on what the company sells and roughly how big it is, from a company record or the company's own site. A title means little until you know whether it was held at a firm of thirty or three thousand.

  4. 4

    Look for public code

    free

    For technical candidates, the name, employer and location are matched to a likely GitHub account with a confidence grade. A confident match gets a summary of repositories, languages and recent activity. These lookups are free, and a weak match is listed as unconfirmed and not used.

  5. 5

    Find the site, talks and writing

    credits

    A web search under the candidate's name and employer looks for conference talks, podcasts and articles, and a personal site is read where the profile or GitHub lists one. Each search is billed whether or not it finds anything, so it runs a small, fixed number of times.

  6. 6

    Read their recent public posts

    credits

    One request returns a page of the candidate's recent public LinkedIn posts. Reposts arrive under the original author and are not attributed to the candidate. Only posts about their work or field go into the brief.

  7. 7

    Write the page and the three questions

    credits

    A metered AI step assembles the page with a source beside each line and the gaps listed as gaps. It then writes the three questions the material raises and does not answer: a short tenure, a change of direction, a title whose scope is unclear.

What a run costs

Credits are spent per candidate the play actually works, and a lookup that finds nothing usually costs nothing. The two figures are the run where the first provider answers and the run where every lookup walks its full chain.

candidatesif the first provider answersif every lookup walks the chain
25$7.171 credits$9.191 credits
100$29290 credits$37370 credits
500$1431430 credits$1831830 credits
1,000$2852850 credits$3653650 credits

Free before anything is charged

  • Confirm who the brief is about
  • Look for public code

What moves the number

  • Coverage on work history. The chain stops at the first provider that answers, and only that provider bills.
  • How many candidates survive the free filters. Everything dropped before the paid steps costs nothing.
  • The scoring and drafting steps run on the metered agent, charged on what they read and write rather than per candidate, so they sit outside this table.
  • Providers you connect with your own key. Those calls bill your account, not your credits.

An estimate, not a quote, priced at the volume credit rate. Your agent sizes the run against your own workspace and tells you what it will cost before it spends anything.

The first call is decided by the questions you ask

A candidate can tell early in a call whether you have read their background or skimmed a headline. The recruiter who asks why they left a platform team after a year for a much smaller company gets a real conversation. The one who asks them to walk through their CV gets the rehearsed version.

The preparation is not difficult, it is scattered. Dated history in one place, what the employers do in another, code on GitHub, a talk on a conference site, a few public posts. This play gathers it onto one page and then does the more useful thing: it works out what the page does not explain. Those gaps are your three questions.

It also draws a line and keeps to it. The brief covers the professional footprint, meaning work the candidate chose to publish about their work. It does not cover age, health, family, beliefs or anything else protected, and it does not read personal social accounts, even where a profile links to one. A brief that strays there is a liability to you and unfair to them.

What you get back

One page in a fixed order. Dated roles with tenure. A line on each recent employer: what it does and roughly how big it is. Public work, each item with its link: repositories and the languages in them, a personal site, talks or articles, recent public posts about their field. Then the three questions, each tied to a fact above. At the foot, what was looked for and not found, any unconfirmed matches, and the list of sources.

Nothing is sent to the candidate and nothing is written to your ATS.

Variations worth knowing

Engineers. Where code is the evidence, the GitHub section can be checked against the tools the role names. Source engineers on proof of work instead of job titles covers that in depth.

History only. For a quick screen, ask for the dated history and the employer lines and skip the rest. It is the lightest version and often enough.

After the call. If the candidate is strong and wrong for this role, score one candidate against every open role in your ATS is the next move.

Where this goes wrong

The wrong person. A common name matches several GitHub accounts and several speakers. Matches are graded, and anything short of confident is kept out of the facts.

A thin footprint is normal. Most good people publish little. An empty public-work section says nothing about ability, and the brief writes "not found" and does not pad.

Reposts are not opinions. A post feed shows reposts under their original author. Only posts the candidate wrote are quoted.

Profiles lag. The history is what the person last wrote down. Confirm the current role on the call.

Treating the brief as an assessment. It prepares a conversation. The questions are questions, not findings.

Questions

Why professional footprint only?

Because that is what a hiring conversation is about, and because the rest is a liability. Age, health, family, beliefs and the like have no place in a recruitment decision, and a brief that collects them puts you on the wrong side of discrimination and data-protection rules whatever you meant to do with it. So the prompt rules them out, personal social accounts are not read even where a profile links to one, and the brief holds only work the candidate chose to publish about their work.

What if the GitHub account or the website is not theirs?

Then it must not be in the brief. A common name matches several accounts and several conference speakers. The GitHub match is graded on name, employer, location and any linked profile, and anything short of confident is listed at the foot as unconfirmed and kept out of the facts. It is better to ask the candidate for their GitHub on the call than to quiz them about somebody else's code.

What are the three questions?

The things the page cannot explain. Why a role lasted under a year. What a head-of title meant in a company that size. Whether the move from a large employer to a small one was a choice. Each question names the fact it comes from, so you can see why it is being asked. If the material only supports two, you get two.

Does it work for candidates who are not engineers?

Yes, with a shorter page. The code section is skipped, and what remains is the dated history, the employer lines, any talks or articles, and recent public posts. Many strong candidates publish nothing at all. The brief says not found and moves on, and an empty public-work section is never read as a mark against anyone.

Will it score the candidate for me?

No. It prepares a conversation and does not assess anyone. Scoring against a spec is a different job, done against must-haves you define, and mixing the two would turn research notes into a verdict the candidate never had a chance to answer. Nothing is written to your ATS and the candidate is not contacted.