trevhud/rote / examples/holdout2/hiring-screener/skill
hiring-screener
VerifiedTakes a pile of applications and turns it into a ranked shortlist, scored only against the job's stated rubric, with drafted replies to every candidate, scheduled interviews, and an onboarding checklist for whoever gets hired. Picks up where job-post-builder leaves off and carries the funnel through to first day, including the payroll setup handoff. Runs on Gmail and Google Calendar, deepens with Gusto, DocuSign, and Drive or M365, and works entirely from uploaded resumes with drafted replies the owner sends by hand. Use this whenever the owner has applicants to deal with — including phrasings like "I've got 40 resumes and no time," "who should I interview," "screen these applications," "help me narrow this down," "set up interviews for the top three," or "I need to tell the rest no." Reach for it when the owner mentions applicants, resumes, candidates, or being buried in a hiring inbox.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 10.0 KB |
| reference/candidate_comms.md | markdown | 4.8 KB |
| reference/fair_screening.md | markdown | 5.4 KB |
| reference/gotchas.md | markdown | 5.5 KB |
| reference/onboarding.md | markdown | 4.3 KB |
| reference/rubric.md | markdown | 5.3 KB |
| sample-input.json | config | 2.4 KB |
reference/rubric.md
View on GitHub# The rubric
The list of things the job actually requires. Everything downstream is measured against it and nothing else.
---
## Where it comes from
**In order of preference:**
1. **From `job-post-builder`.** That skill produces the post, the structured interview guide, and the scoring rubric together. If it ran, the rubric already exists and is already agreed.
2. **Derived from the job post.** Read the post and pull out every stated requirement, responsibility, and preference. Show it back for correction.
3. **Built with the owner now.** Fifteen minutes of questions. Use this only when there is no post.
**Do not read resumes before the rubric exists.** Once you have seen the candidates, the rubric gets written to fit whoever looked good, which is exactly backwards and impossible to defend later.
---
## What goes in it
Three tiers, and the tier matters more than the wording.
**Must-have.** Absence disqualifies. Keep this list short and honest — three or four items. A must-have that is really a nice-to-have shrinks the field for no reason and disproportionately screens out good candidates from non-traditional paths.
Examples: the license the work legally requires, a clean driving record for a driving role, availability for the actual shift.
**Should-have.** The real differentiators. Weighted, scored, and where the ranking comes from.
Examples: years doing comparable work, specific systems or equipment, scale handled, supervisory experience where the role includes it.
**Nice-to-have.** Small positive weight. Never disqualifying.
---
## Weighting
Assign weights to the should-have criteria totaling 100. Force the owner to make the trade-off explicitly, because they are making it either way.
Ask directly: "If one candidate has five years of commercial rooftop work and no supervisory experience, and another has two years plus three years leading a crew — who do you want in the room?" The answer sets the weights faster than any abstract discussion.
**Cap any single criterion at 40.** Above that the rubric is really a one-question screen, and everything else is decoration.
---
## Writing a criterion so it can be scored
A criterion has to be checkable against a resume. Vague criteria get scored on impressions.
**Bad:** "Strong communicator." Unmeasurable from a resume, and in practice it becomes a proxy for writing polish, which is a proxy for schooling.
**Good:** "Has worked directly with commercial customers on site — scheduling, explaining work, handling complaints." Now there is something to look for and quote.
**Bad:** "Culture fit." Not a criterion. It is the place unexamined preference hides, and it will not survive being read aloud.
**Good:** Whatever the owner actually means by it, stated as behavior: "Has worked in a small team without a dispatcher, self-directed on route."
---
## The scale
Four points per criterion. More granularity is false precision from a one-page document.
| Score | Means |
|---|---|
| 3 | Clear evidence, at or above what the role needs |
| 2 | Evidence of something close or adjacent |
| 1 | Little evidence, but not contradicted |
| 0 | Absent, or evidence points the other way |
| — | Cannot assess from what was submitted |
**The dash is not a zero.** It routes the candidate to the "cannot assess" band and generates a question, not a rejection.
---
## The scoring math
One formula, used for every candidate:
**weighted = sum(score ÷ 3 × weight) over scored criteria; dashes excluded from numerator and denominator; report as % of assessed weight.**
Read that as three rules:
1. **Each scored criterion contributes score ÷ 3 of its weight.** A 3 earns the full weight, a 2 earns two-thirds, a 1 earns one-third, a 0 earns none.
2. **A dash is excluded from both sides.** It does not earn points and its weight drops out of the total the candidate is measured against. That is what keeps a missing fact from acting like a zero.
3. **Report the result as a percentage of the weight actually assessed**, and say how much weight that was.
Worked through, for a candidate scoring 3 on rooftop (weight 35), 3 on diagnostics (25), 2 on customer contact (20), and dashes on work orders (15) and controls (5):
```
35 × 3/3 = 35.0
25 × 3/3 = 25.0
20 × 2/3 = 13.3
73.3 points earned
Assessed weight 80 (100 less the 20 behind the two dashes)
Reported score 73.3 ÷ 80 = 92
```
Always show the assessed weight next to the score. 92 out of 80 assessed weight is a very different candidate from 92 out of 100, and the difference is exactly the questions worth asking them.
---
## Worked example — Okonkwo Mechanical, Commercial Service Tech
```
MUST-HAVE
EPA 608 Universal certification
Valid driver's license, insurable record
Available weekdays 7am-4pm, on-call rotation 1 week in 4
SHOULD-HAVE (weighted to 100)
35 Commercial rooftop unit experience — install, service, or both
25 Diagnostic depth on packaged units and splits
20 Direct customer contact on site
15 Documented work orders / service records
5 Controls and BAS familiarity
NICE-TO-HAVE
Refrigeration experience beyond HVAC
Spanish
```
Note what is not there: no degree, no "professional appearance," no minimum years figure that would exclude a strong two-year tech. Everything listed is something Ray would defend out loud to the person he did not hire.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |