Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Point Cowork at a folder of resumes plus a job description -- screens every candidate against the actual requirements, produces a ranked shortlist with evidence, drafts advance/decline emails, and builds interview kits for the top picks. Pairs with hiring-scorecard for the interview stage.
.claude/skills/onewave-ai-cowork-hiring-screener/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 198% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 101% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 66% | 0% |
| case-21 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-22 | ✗→✓ | ▲ Improved | -3% | 0% |
Screen a resume pile the way a disciplined recruiter does: score against the written requirements, cite evidence from the resume for every score, and never let formatting quality masquerade as candidate quality. Input is a folder of resumes (PDF, .docx, text) and a job description; output is a defensible shortlist.
screening-report.md: Tier 1 (interview now), Tier 2 (backup), Tier 3 (decline), each candidate with score breakdown, one-paragraph summary, strongest signal, and biggest gap or open question.hiring-scorecard for structured interview evaluation.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→fail | 34,897 | 24,017 | -31% | 1 | 1 | 0% | 5,133 | 3,548 | -31% | 0 | 0 | — |
case-02 | fail→fail | 6,157 | 9,921 | +61% | 1 | 1 | 0% | 373 | 1,415 | +279% | 0 | 0 | — |
case-03 | fail→fail | 6,612 | 7,992 | +21% | 1 | 1 | 0% | 966 | 1,689 | +75% | 0 | 0 | — |
case-04 | pass→pass | 9,521 | 8,219 | -14% | 1 | 1 | 0% | 1,533 | 1,947 | +27% | 0 | 0 | — |
case-05 | pass→pass | 17,274 | 17,804 | +3% | 1 | 1 | 0% | 2,642 | 3,105 | +18% | 0 | 0 | — |
case-06 | fail→pass | 4,969 | 10,318 | +108% | 1 | 1 | 0% | 733 | 2,186 | +198% | 0 | 0 | — |
case-07 | pass→fail | 7,891 | 2,954 | -63% | 1 | 1 | 0% | 1,215 | 1,006 | -17% | 0 | 0 | — |
case-08 | pass→pass | 9,070 | 6,304 | -30% | 1 | 1 | 0% | 1,497 | 1,675 | +12% | 0 | 0 | — |
case-09 | fail→fail | 9,977 | 10,618 | +6% | 1 | 1 | 0% | 1,930 | 2,531 | +31% | 0 | 0 | — |
case-10 | pass→pass | 10,280 | 11,288 | +10% | 1 | 1 | 0% | 1,517 | 2,385 | +57% | 0 | 0 | — |
case-11 | pass→pass | 6,048 | 10,152 | +68% | 1 | 1 | 0% | 1,008 | 2,269 | +125% | 0 | 0 | — |
case-12 | pass→pass | 10,246 | 6,492 | -37% | 1 | 1 | 0% | 1,564 | 1,657 | +6% | 0 | 0 | — |
case-13 | pass→pass | 8,806 | 6,193 | -30% | 1 | 1 | 0% | 1,293 | 1,665 | +29% | 0 | 0 | — |
case-14 | pass→pass | 13,850 | 7,378 | -47% | 1 | 1 | 0% | 1,992 | 1,732 | -13% | 0 | 0 | — |
case-15 | pass→pass | 19,066 | 14,645 | -23% | 1 | 1 | 0% | 2,725 | 2,748 | +1% | 0 | 0 | — |
case-16 | pass→pass | 11,308 | 8,513 | -25% | 1 | 1 | 0% | 1,722 | 1,965 | +14% | 0 | 0 | — |
case-17 | fail→fail | 3,652 | 5,086 | +39% | 1 | 1 | 0% | 512 | 904 | +77% | 0 | 0 | — |
case-18 | pass→fail | 10,263 | 4,427 | -57% | 1 | 1 | 0% | 1,914 | 1,091 | -43% | 0 | 0 | — |
case-19 | fail→pass | 5,709 | 8,145 | +43% | 1 | 1 | 0% | 861 | 1,734 | +101% | 0 | 0 | — |
case-20 | fail→pass | 5,389 | 4,940 | -8% | 1 | 1 | 0% | 817 | 1,354 | +66% | 0 | 0 | — |
case-21 | fail→pass | 11,679 | 5,518 | -53% | 1 | 1 | 0% | 1,764 | 1,473 | -16% | 0 | 0 | — |
case-22 | fail→pass | 8,316 | 4,133 | -50% | 1 | 1 | 0% | 1,288 | 1,251 | -3% | 0 | 0 | — |
case-23 | pass→pass | 16,263 | 9,122 | -44% | 1 | 1 | 0% | 2,429 | 2,051 | -16% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted, and 22 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +9 percentage points is the difference between those two pass rates over the 22 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.