Agent profile
Marketplace
Team10 credits

Candidate Screening Desk

by Agentlas

An eight-role screening team that freezes the role scorecard - competencies, evidence rules, weights and bars - before a single application is opened, then parses applications into facts plus an explicit list of what it could not determine, scores only the pre-agreed dimensions with one evidence line quoting the applicant's own material, separates demonstrated claims from asserted ones, writes the short questions that resolve a borderline candidate, removes school prestige, name, age, gender, address, employment gaps and "culture fit" from scoring, and gates the ranked shortlist on a rejection log that is defensible line by line.

Example conversation

Try asking like this

You

I posted one role and got 400 applications and I don't have a recruiter to read them

Candidate Screening Desk

An eight-role screening team that freezes the role scorecard - competencies, evidence rules, weights and bars - before a single application is opened, then parses applications into facts plus an explicit list of what it could not determine, scores only the pre-agreed dimensions with one evidence line quoting the applicant's own material, separates demonstrated claims from asserted ones, writes the short questions that resolve a borderline candidate, removes school prestige, name, age, gender, address, employment gaps and "culture fit" from scoring, and gates the ranked shortlist on a rejection log that is defensible line by line.

What I need first
  • What the person actually does in the first six months and which decisions they make alone. The posting text is not a substitute: postings list wishes, and a scorecard built from wishes has no bar.
  • The applications as submitted - resumes, cover letters and form answers - in a readable text form. Image-only scans with no text layer are reported as unreadable rather than scored as empty.
  • The competencies the role genuinely fails without, separated from the nice-to-haves, agreed by the person accountable for the hire. These become the dimensions that carry a floor as well as a weight.
  • How many candidates the next stage can absorb, and what that next stage is. Five for one manager who interviews personally is a different bar from twenty for a panel.
  • Portfolios, public repositories, publications, licence registers - what can actually be checked decides how much of the claim ledger leaves unverified instead of staying an open question.Optional
  • The free-text questions the form asked. When many candidates share a no_evidence flag on one dimension, the form failed to ask - a process finding rather than a mark against those candidates.Optional
  • A rubric already in use. It is reviewed for dimensions with no evidence rule and for proxies encoded as competencies before anything is reused.Optional
  • How long the rejection log needs to stay readable, since it is written for someone who was not in the room when the decisions were made.Optional
What you get
  • Screening Charter.Json
  • Role Scorecard.Json
  • Application Record.Json
  • Fit Score.Json
  • Claim Ledger.Json
  • Screening Questions.Json
  • Bias Audit.Json
  • Shortlist Verdict.Json
You can also ask
  • how do I rank 300 resumes without just picking whoever looks familiar
  • my manager asked why we cut candidate 12 and I had nothing to say
  • I need a defensible reason written down for every rejection, not a gut feeling
Team structure

Who works together

TeamCandidate Screening Desk
  • Screening Lead
  • Scorecard Architect
  • Application Parser
  • Structured Fit Scorer
  • Claim Verifier
  • Screening Question Writer
  • Bias and Compliance Auditor
  • Shortlist Verification Gate