Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →- Conduct performance reviews for employees each quarter (after 3 tasks completed)
.claude/skills/1mancompany-reviews-performance-review-capability/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 391% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 323% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 71% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 1% | 0% |
Only three tiers are allowed; no other scores may be used:
| Score | Tier | Meaning | |-------|------|---------| | 3.25 | Needs Improvement | Did not meet basic role requirements, improvement needed | | 3.50 | Satisfactory | Met role requirements, stable performance | | 3.75 | Excellent | Exceeded expectations, outstanding performance |
Promotion criteria: 3 consecutive quarters of 3.75 (Excellent) → automatic promotion by one level
Regular employee level cap: Lv.3 (Senior)
| Level | Prefix | Example Titles | |-------|--------|---------------| | Lv.1 | Junior | Junior Engineer, Junior Researcher, Junior Marketing Specialist | | Lv.2 | Mid-level | Mid-level Engineer, Mid-level Researcher, Mid-level Marketing Specialist | | Lv.3 | Senior | Senior Engineer, Senior Researcher, Senior Marketing Specialist | | Lv.4 | Founding | Founding employee (cannot be achieved through promotion) | | Lv.5 | CEO | CEO (cannot be achieved through promotion) |
Note: Lv.4 (Founding) and Lv.5 (CEO) are not part of the performance promotion system.
Review results must include JSON format:
json{"action": "review", "reviews": [{"id": "employee_id", "score": 3.5, "feedback": "..."}]}
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 9,711 | 8,762 | -10% | 1 | 1 | 0% | 433 | 2,125 | +391% | 0 | 0 | — |
case-02 | fail→pass | 13,418 | 6,191 | -54% | 1 | 1 | 0% | 361 | 1,528 | +323% | 0 | 0 | — |
case-03 | fail→pass | 7,508 | 6,105 | -19% | 1 | 1 | 0% | 1,355 | 1,604 | +18% | 0 | 0 | — |
case-04 | fail→pass | 5,018 | 6,428 | +28% | 1 | 1 | 0% | 931 | 1,591 | +71% | 0 | 0 | — |
case-05 | fail→pass | 6,055 | 3,078 | -49% | 1 | 1 | 0% | 985 | 991 | +1% | 0 | 0 | — |
case-06 | fail→pass | 6,233 | 4,350 | -30% | 1 | 1 | 0% | 1,176 | 1,245 | +6% | 0 | 0 | — |
case-17 | fail→pass | 6,885 | 5,825 | -15% | 1 | 1 | 0% | 1,242 | 1,559 | +26% | 0 | 0 | — |
case-07 | fail→pass | 5,768 | 3,961 | -31% | 1 | 1 | 0% | 1,127 | 1,234 | +9% | 0 | 0 | — |
case-08 | fail→pass | 4,975 | 4,051 | -19% | 1 | 1 | 0% | 908 | 1,172 | +29% | 0 | 0 | — |
case-09 | fail→pass | 6,592 | 3,836 | -42% | 1 | 1 | 0% | 1,280 | 1,072 | -16% | 0 | 0 | — |
case-10 | fail→pass | 5,733 | 3,849 | -33% | 1 | 1 | 0% | 994 | 1,088 | +9% | 0 | 0 | — |
case-16 | fail→pass | 3,610 | 3,552 | -2% | 1 | 1 | 0% | 712 | 1,157 | +63% | 0 | 0 | — |
case-11 | fail→pass | 9,021 | 6,954 | -23% | 1 | 1 | 0% | 1,601 | 1,871 | +17% | 0 | 0 | — |
case-12 | fail→pass | 6,635 | 4,034 | -39% | 1 | 1 | 0% | 1,250 | 1,168 | -7% | 0 | 0 | — |
case-13 | pass→pass | 6,445 | 4,301 | -33% | 1 | 1 | 0% | 1,283 | 1,268 | -1% | 0 | 0 | — |
case-14 | pass→fail | 12,201 | 8,715 | -29% | 1 | 1 | 0% | 2,183 | 2,051 | -6% | 0 | 0 | — |
case-15 | fail→pass | 4,224 | 5,315 | +26% | 1 | 1 | 0% | 719 | 1,399 | +95% | 0 | 0 | — |
case-18 | pass→pass | 5,855 | 3,651 | -38% | 1 | 1 | 0% | 1,081 | 1,025 | -5% | 0 | 0 | — |
case-19 | fail→pass | 6,873 | 4,802 | -30% | 1 | 1 | 0% | 1,304 | 1,443 | +11% | 0 | 0 | — |
case-20 | pass→pass | 12,441 | 11,152 | -10% | 1 | 1 | 0% | 2,226 | 2,426 | +9% | 0 | 0 | — |
case-21 | pass→pass | 2,491 | 2,794 | +12% | 1 | 1 | 0% | 470 | 938 | +100% | 0 | 0 | — |
case-22 | pass→pass | 14,192 | 10,291 | -27% | 1 | 1 | 0% | 2,447 | 2,102 | -14% | 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. 22 cases were attempted, and 20 counted toward the lift figure. The other 2 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 +68 percentage points is the difference between those two pass rates over the 20 comparable cases. 1 case got worse with the skill loaded, and it is 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.