Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Runs a relentless interview that sharpens a plan or design and creates docs (ADRs and glossary) along the way. Use when the user wants to stress-test a plan or design, or produce decision records.
.claude/skills/fradser-grill-with-docs/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-20 | ✓→✗ | ▼ Worse | 148% | 0% |
| case-21 | ✓→✓ | ＝ Same ✓ | 88% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | 20% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | -37% | 0% |
| case-02 | ✗→✗ | ＝ Same ✗ | 27% | 0% |
Run a /mattpocock:grilling session, using the /mattpocock:domain-modeling skill.
Run the /mattpocock:grilling session with the /mattpocock:domain-modeling skill active. Every sharpened term and locked decision lands in CONTEXT.md or an ADR as it crystallises — the paper trail is what distinguishes this skill from /mattpocock:grill-me. Interview questions go through the AskUserQuestion tool, one per call, as /mattpocock:grilling dictates.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 4,968 | 6,684 | +35% | 1 | 1 | 0% | 668 | 422 | -37% | 0 | 0 | — |
case-02 | fail→fail | 6,259 | 9,614 | +54% | 1 | 1 | 0% | 1,032 | 1,309 | +27% | 0 | 0 | — |
case-03 | fail→fail | 6,238 | 6,353 | +2% | 1 | 1 | 0% | 955 | 414 | -57% | 0 | 0 | — |
case-04 | fail→fail | 7,307 | 9,882 | +35% | 1 | 1 | 0% | 992 | 929 | -6% | 0 | 0 | — |
case-05 | fail→fail | 4,477 | 5,704 | +27% | 1 | 1 | 0% | 688 | 448 | -35% | 0 | 0 | — |
case-06 | fail→fail | 7,583 | 13,935 | +84% | 1 | 1 | 0% | 1,110 | 2,406 | +117% | 0 | 0 | — |
case-07 | fail→fail | 4,154 | 5,578 | +34% | 1 | 1 | 0% | 575 | 501 | -13% | 0 | 0 | — |
case-08 | fail→fail | 6,599 | 5,920 | -10% | 1 | 1 | 0% | 955 | 393 | -59% | 0 | 0 | — |
case-09 | fail→fail | 9,869 | 5,121 | -48% | 1 | 1 | 0% | 1,491 | 544 | -64% | 0 | 0 | — |
case-10 | fail→fail | 4,544 | 11,548 | +154% | 1 | 1 | 0% | 576 | 1,887 | +228% | 0 | 0 | — |
case-11 | fail→fail | 7,214 | 15,181 | +110% | 1 | 1 | 0% | 1,106 | 2,541 | +130% | 0 | 0 | — |
case-12 | fail→fail | 12,147 | 7,404 | -39% | 1 | 1 | 0% | 1,754 | 563 | -68% | 0 | 0 | — |
case-13 | fail→fail | 9,040 | 8,340 | -8% | 1 | 1 | 0% | 1,339 | 782 | -42% | 0 | 0 | — |
case-19 | fail→fail | 5,935 | 4,102 | -31% | 1 | 1 | 0% | 1,046 | 325 | -69% | 0 | 0 | — |
case-14 | fail→fail | 4,347 | 9,947 | +129% | 1 | 1 | 0% | 651 | 1,457 | +124% | 0 | 0 | — |
case-15 | fail→fail | 4,228 | 5,915 | +40% | 1 | 1 | 0% | 586 | 501 | -15% | 0 | 0 | — |
case-16 | fail→fail | 4,804 | 10,345 | +115% | 1 | 1 | 0% | 566 | 924 | +63% | 0 | 0 | — |
case-17 | fail→fail | 7,562 | 11,224 | +48% | 1 | 1 | 0% | 1,088 | 1,174 | +8% | 0 | 0 | — |
case-18 | fail→fail | 4,405 | 3,835 | -13% | 1 | 1 | 0% | 576 | 340 | -41% | 0 | 0 | — |
case-20 | pass→fail | 7,857 | 32,849 | +318% | 1 | 1 | 0% | 1,468 | 3,639 | +148% | 0 | 0 | — |
case-21 | pass→pass | 5,637 | 9,156 | +62% | 1 | 1 | 0% | 855 | 1,604 | +88% | 0 | 0 | — |
case-22 | pass→pass | 6,773 | 10,449 | +54% | 1 | 1 | 0% | 1,477 | 1,777 | +20% | 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 8 counted toward the lift figure. The other 14 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 -5 percentage points is the difference between those two pass rates over the 8 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Other measured skills in the registry, with their headline benchmark lift.