# Hiring And Interviewing

> Designs and runs hiring — role definition, sourcing, interview loop design, structured evaluation, and the decision itself. Use this to open a role, write a job description or scorecard, design an interview process, prepare interview questions, calibrate a hiring decision, or diagnose why a hiring process produces poor outcomes.

- Skill: `cbrock84/hiring-and-interviewing` (Agent Skill)
- Install (CLI): `npx skillmds@latest add cbrock84/hiring-and-interviewing`
- Raw SKILL.md: https://api.skillmd.com/api/skills/cbrock84/hiring-and-interviewing/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: cbrock84 (https://skillmd.com/u/cbrock84)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/cbrock84/hiring-and-interviewing

---


# Hiring and interviewing

## Define before sourcing

Write, before posting anything:

- **What this person will accomplish in their first year** — outcomes, not responsibilities.
- **The three or four competencies** that actually predict those outcomes.
- **What is genuinely required** versus what is preferred. Long requirement lists are the most
  reliable way to shrink and homogenize a candidate pool, and most of the list is never used in the
  decision.
- **The bar**, agreed by everyone on the loop, before the first interview.

Undefined roles produce interview loops where each interviewer evaluates against a private
definition, and the decision goes to whoever argues hardest.

## Design the loop

Each interview assesses different competencies, stated in advance. Overlapping interviews produce
four opinions on the same thing and none on the rest.

- **Structured beats unstructured**, consistently and by a wide margin. Same questions, same order,
  same rubric. Unstructured interviews mostly measure similarity to the interviewer.
- **Work samples predict best.** A realistic exercise close to the actual job beats any amount of
  discussion about it. Keep it time-boxed and pay for anything substantial.
- **Behavioral questions about the past**, with follow-ups for specifics: what was the situation,
  what did *you* do, what happened, what would you change. Hypotheticals measure articulacy.
- Fewer, better interviews. Long loops lose good candidates and add little signal after the fourth
  conversation.

## Evaluating

Interviewers write their assessment against the rubric **before** any discussion. Group discussion
first produces convergence on the first confident opinion rather than an aggregation of independent
ones.

Debrief on evidence: what did they say or do that supports this rating? "Culture fit" without
behavioral evidence is where bias enters, and it should be challenged every time it appears.

## Deciding

A yes needs evidence on every required competency, not a strong overall impression. Where evidence
is missing, get it — an extra conversation is cheap compared to a mis-hire.

Ambiguity means no. The cost of a bad hire is far larger and lasts far longer than the cost of a
longer search, and it is borne by the team, not the hiring manager.

## Tooling

Applicant tracking, roughly by scale: Workable, JazzHR or the hiring module of an all-in-one HRIS
at small scale; Greenhouse, Lever or Ashby once structured interviewing and reporting matter;
Workday Recruiting or SmartRecruiters at enterprise scale, and similar.

Sourcing and scheduling: LinkedIn Recruiter, Gem, GoodTime, and similar.

Put the scorecard in the tracking system rather than beside it. A structured process that lives in
a separate document is a structured process that stops being followed in the first busy week.

## Never

- Lower the bar because the search has been long. Reopen the role definition instead.
- Let one interviewer's strong view override written independent assessments.
- Skip reference checks on judgment and collaboration, which are exactly what interviews measure
  worst.