Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Explains a single control once and shows every framework it maps to via the SCF crosswalk. Resolves SCF IDs, framework-specific IDs, and plain-English descriptions. Never reproduces normative text.
.claude/skills/grcengclub-control-explainer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 9% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -6% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 26% | 0% |
You are the skill invoked by /teach-me:control <control-id>. Your job is to take any reasonable reference to a control — SCF ID, framework-specific ID, or English description — and produce a single explanation that is useful across every framework that maps to it.
CC6.1; NIST 800-53 calls it IA-2; ISO 27001 calls it 8.3. Show the learner that the underlying requirement is shared — that's the point of the SCF crosswalk./grc-engineer:test-control to validate end-to-end.IAC-01, CHG-02), use it directly.CC6.1, IA-2, 8.3, Req 7.2), call /grc-engineer:map-controls-unified --framework=<best-guess> to map it to SCF first. If --lens=<framework> is provided, use that as the lens./grc-engineer:map-controls-unified <scf-id> to get every framework that maps this SCF control, with their local IDs.plugins/frameworks/, read skills/<framework>-expert/SKILL.md for any framework-specific notes the plugin author left about this control./grc-engineer:test-control.Markdown. Use a small table for the cross-framework view. Keep prose paragraphs short. Bold the section headings.
/grc-engineer:generate-implementation for that.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | fail→fail | 31,118 | 14,294 | -54% | 1 | 1 | 0% | 5,718 | 3,152 | -45% | 0 | 0 | — |
case-01 | fail→fail | 25,646 | 20,939 | -18% | 1 | 1 | 0% | 3,778 | 2,731 | -28% | 0 | 0 | — |
case-02 | fail→pass | 18,085 | 11,512 | -36% | 1 | 1 | 0% | 2,758 | 2,654 | -4% | 0 | 0 | — |
case-03 | fail→fail | 19,313 | 11,001 | -43% | 1 | 1 | 0% | 3,235 | 2,591 | -20% | 0 | 0 | — |
case-05 | fail→fail | 19,768 | 13,217 | -33% | 1 | 1 | 0% | 3,504 | 3,110 | -11% | 0 | 0 | — |
case-06 | fail→pass | 12,384 | 9,940 | -20% | 1 | 1 | 0% | 2,175 | 2,583 | +19% | 0 | 0 | — |
case-07 | pass→pass | 7,633 | 9,950 | +30% | 1 | 1 | 0% | 1,204 | 2,402 | +100% | 0 | 0 | — |
case-08 | fail→pass | 14,407 | 10,402 | -28% | 1 | 1 | 0% | 2,358 | 2,571 | +9% | 0 | 0 | — |
case-09 | fail→pass | 18,920 | 13,627 | -28% | 1 | 1 | 0% | 3,101 | 2,930 | -6% | 0 | 0 | — |
case-10 | fail→pass | 16,146 | 14,182 | -12% | 1 | 1 | 0% | 2,571 | 3,237 | +26% | 0 | 0 | — |
case-11 | fail→pass | 23,371 | 11,117 | -52% | 1 | 1 | 0% | 4,150 | 2,698 | -35% | 0 | 0 | — |
case-12 | fail→pass | 13,323 | 11,560 | -13% | 1 | 1 | 0% | 2,074 | 2,926 | +41% | 0 | 0 | — |
case-13 | fail→pass | 15,231 | 13,182 | -13% | 1 | 1 | 0% | 2,459 | 3,143 | +28% | 0 | 0 | — |
case-14 | pass→pass | 15,664 | 13,118 | -16% | 1 | 1 | 0% | 2,357 | 2,982 | +27% | 0 | 0 | — |
case-15 | pass→pass | 12,440 | 15,604 | +25% | 1 | 1 | 0% | 2,074 | 3,402 | +64% | 0 | 0 | — |
case-16 | fail→pass | 17,928 | 11,163 | -38% | 1 | 1 | 0% | 2,703 | 2,541 | -6% | 0 | 0 | — |
case-17 | fail→pass | 17,429 | 16,790 | -4% | 1 | 1 | 0% | 2,546 | 3,466 | +36% | 0 | 0 | — |
case-18 | fail→pass | 15,830 | 9,631 | -39% | 1 | 1 | 0% | 2,315 | 2,353 | +2% | 0 | 0 | — |
case-19 | fail→pass | 19,127 | 12,501 | -35% | 1 | 1 | 0% | 3,323 | 2,771 | -17% | 0 | 0 | — |
case-20 | fail→pass | 22,800 | 9,775 | -57% | 1 | 1 | 0% | 4,169 | 2,404 | -42% | 0 | 0 | — |
case-21 | fail→pass | 9,734 | 17,291 | +78% | 1 | 1 | 0% | 1,528 | 3,552 | +132% | 0 | 0 | — |
case-22 | fail→fail | 28,093 | 23,159 | -18% | 1 | 1 | 0% | 6,174 | 5,321 | -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. The headline lift of +64 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.