Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Deep Seyfert v5 guidance for building, updating, debugging, reviewing, and testing Discord bots that use the Seyfert TypeScript framework. Use when a task mentions Seyfert, imports from the `seyfert` package, `seyfert.config` files, Seyfert class/decorator APIs (`Command`, `Declare`, `Options`, `ComponentCommand`, `createEvent`, `createPlugin`), Slipher packages, @slipher/testing, or yunaforseyfert.
.claude/skills/tiramisulabs-seyfert/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-20 | ✓→✗ | ▼ Worse | 76% | 0% |
| case-22 | ✓→✗ | ▼ Worse | -10% | 0% |
| case-18 | ✓→✓ | ＝ Same ✓ | -3% | 0% |
Use the target project's installed seyfert declarations and runtime as the authority. This skill's stable baseline is seyfert@5.1.0; read source truth for provenance, version scope, and known docs corrections. Verify separately versioned @slipher/* and other external packages against their installed versions.
@slipher/testing: testingPrefer root imports. Before using a deep import or copying a version-sensitive example, check the installed lib/index.d.ts, relevant declarations, or the corresponding source barrel. Keep existing project conventions and run its focused type checks and tests.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-18 | pass→pass | 25,497 | 22,437 | -12% | 1 | 1 | 0% | 3,462 | 3,366 | -3% | 0 | 0 | — |
case-19 | pass→pass | 15,871 | 23,281 | +47% | 1 | 1 | 0% | 2,607 | 3,059 | +17% | 0 | 0 | — |
case-20 | pass→fail | 18,702 | 23,325 | +25% | 1 | 1 | 0% | 1,953 | 3,439 | +76% | 0 | 0 | — |
case-07 | pass→pass | 17,533 | 14,254 | -19% | 1 | 1 | 0% | 1,314 | 1,837 | +40% | 0 | 0 | — |
case-01 | fail→fail | 21,170 | 26,631 | +26% | 1 | 1 | 0% | 2,905 | 4,311 | +48% | 0 | 0 | — |
case-02 | fail→fail | 14,246 | 13,530 | -5% | 1 | 1 | 0% | 2,689 | 2,762 | +3% | 0 | 0 | — |
case-03 | pass→pass | 10,996 | 36,419 | +231% | 1 | 1 | 0% | 1,822 | 1,363 | -25% | 0 | 0 | — |
case-04 | pass→pass | 16,668 | 10,024 | -40% | 1 | 1 | 0% | 2,101 | 1,980 | -6% | 0 | 0 | — |
case-05 | pass→pass | 53,508 | 22,910 | -57% | 1 | 1 | 0% | 4,214 | 3,805 | -10% | 0 | 0 | — |
case-06 | pass→pass | 5,611 | 4,221 | -25% | 1 | 1 | 0% | 548 | 903 | +65% | 0 | 0 | — |
case-08 | pass→pass | 8,418 | 13,682 | +63% | 1 | 1 | 0% | 1,176 | 1,715 | +46% | 0 | 0 | — |
case-09 | fail→pass | 12,512 | 12,446 | -1% | 1 | 1 | 0% | 1,221 | 1,362 | +12% | 0 | 0 | — |
case-10 | pass→pass | 18,274 | 16,787 | -8% | 1 | 1 | 0% | 2,147 | 2,452 | +14% | 0 | 0 | — |
case-11 | pass→pass | 17,559 | 10,972 | -38% | 1 | 1 | 0% | 1,817 | 1,783 | -2% | 0 | 0 | — |
case-12 | fail→pass | 13,272 | 8,160 | -39% | 1 | 1 | 0% | 1,274 | 1,093 | -14% | 0 | 0 | — |
case-13 | pass→pass | 13,706 | 23,817 | +74% | 1 | 1 | 0% | 1,518 | 1,221 | -20% | 0 | 0 | — |
case-14 | pass→pass | 11,532 | 21,424 | +86% | 1 | 1 | 0% | 1,973 | 3,079 | +56% | 0 | 0 | — |
case-15 | pass→pass | 20,526 | 18,146 | -12% | 1 | 1 | 0% | 2,769 | 3,476 | +26% | 0 | 0 | — |
case-16 | pass→pass | 15,942 | 7,987 | -50% | 1 | 1 | 0% | 1,884 | 1,325 | -30% | 0 | 0 | — |
case-17 | pass→pass | 18,292 | 26,329 | +44% | 1 | 1 | 0% | 3,166 | 4,233 | +34% | 0 | 0 | — |
case-21 | pass→pass | 23,067 | 12,685 | -45% | 1 | 1 | 0% | 2,737 | 2,507 | -8% | 0 | 0 | — |
case-22 | pass→fail | 15,606 | 11,322 | -27% | 1 | 1 | 0% | 1,401 | 1,256 | -10% | 0 | 0 | — |
case-23 | pass→pass | 20,411 | 16,540 | -19% | 1 | 1 | 0% | 2,488 | 2,869 | +15% | 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. The headline lift of -33 percentage points is the difference between those two pass rates over the 23 comparable cases. 4 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/31/2026 | +25% |
| gemini-3.6-flash | verified | 8/6/2026 | +41% |
Other measured skills in the registry, with their headline benchmark lift.