---
name: "role-intake-and-scorecard"
description: "Use when a new role opens or a req needs rescoping: agree a checkable role scorecard and hiring plan with the hiring manager before any posting or sourcing."
triggers: ["scope this new role", "kick off a new req", "write a role scorecard", "agree the hiring bar", "intake with the hiring manager", "what should this role require", "calibrate on a benchmark hire"]
version: "1"
---
# Role intake and scorecard
Every search starts here. The role scorecard you write is the bar that
postings, sourcing, screening and interview kits all read.
## What to collect
The hiring manager's ask in any form (call notes, a rough req, an old posting)
and the team context: who the hire reports to and what gap opened the role.
Ask for anything missing one question at a time, and keep drafting.
## Start from the work, not the title
In plain outcomes, write what the hire must own by the end of year one, their
closest collaborators, and the two or three problems the company needs solved.
Quote the hiring manager's sentence behind each point so misreadings surface
now.
## Write the bar
**Must-haves:** three to six, each confirmable from public work or a
structured interview question, so two reviewers with the same evidence would
agree. "Has run on-call for a service with paying customers" passes; "great
ownership mentality" fails until you ask what ownership looks like in practice
and write that. Swap years-of-experience floors and degree requirements for
the outcome behind them unless the hiring manager can defend them.
**Nice-to-haves:** labelled separately; they break ties and never keep anyone
out.
**Disqualifiers:** in the hiring manager's words, then through the fairness
check. Three straight jobs under a year each is a note on the candidate card
for the interviewer, never a disqualifier.
## Keep it fair and job-related
Every line must tie to the work. None may name or proxy for a protected
characteristic such as age, race, ethnicity, national origin, sex, gender
identity, sexual orientation, religion, disability, pregnancy or family
status. "Digital native", "native English speaker", "culture fit" and
graduation-year cut-offs become the real skill underneath ("writes
customer-facing docs in English") or get cut, and you say which.
## Calibrate against one real person
Ask whether someone's work, inside the company or out, sets the bar. If it
does, list four to six checkable things that make that person the benchmark:
keep what transfers (owning a hard problem end to end), drop the incidental
(school, employer brand, exact tools). Take one round of edits; sourcing's
lookalike search runs from this list.
## Map where the people are
- **Target companies:** where this work really happens (stage, size, product),
with 10 to 20 named examples.
- **Boundaries:** whether direct competitors are fair game, and which
companies are off limits (customers, partners, a non-solicit clause).
- **Title variants:** five to ten current and adjacent titles.
- **Evidence sources:** engineers show public code and packages; designers,
portfolios and case studies; marketing and research, articles and talks;
support and solutions, forum answers and help-center writing; sales and
finance, panels, podcasts and published decks.
## Agree the hiring plan in the same sitting
Order stages so deal-breakers are tested first: screen, knockout competency,
deep-dive, hiring-manager close. Name the panel and each interviewer's
competency, the target slate size, the fill date, and who sets compensation
and how. Never propose a pay figure yourself.
You recommend; the hiring manager decides. Nothing reaches a posting, screen
or sourcing run until they approve the scorecard, and later edits go back to
them too.
## What you hand back
The role scorecard in chat (must-haves, nice-to-haves, disqualifiers,
benchmark traits, target companies and boundaries, title variants, evidence
sources), then the hiring plan. After one round of edits, save both to the
hiring folder.
## Fallbacks
No hiring manager to hand? Take the title and ask four questions: what the
hire must deliver in year one, which two things rule a candidate out, whether
anyone's work sets the bar, and when the seat must be filled. Label the result
"draft pending calibration". Sourcing against it is allowed if every batch
flags the bar as uncalibrated.