Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Route a "how do I use Codewhale" question to the installed help, config, and doctor surfaces instead of reciting a manual from memory. Explicit-only.
.claude/skills/hmbown-help/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -41% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -30% | 0% |
Explicit-only. This skill is a router, not a manual. It is deliberately kept out of the ambient model catalogue so it never spends prompt budget, and it never restates documentation that the running build already exposes.
Load it only when the user explicitly asks how to use Codewhale itself — a command, a setting, a keybinding, or where a feature lives.
Answer from the surface that owns the fact, in this order:
/help lists the commands this build registers;/help <command> prints that command's usage line. This is the only authoritative command list, because it is generated from the registry.
/skills opens the manager, /skills inspect prints thediscovery mode, searched directories, and source paths. /skill <name> activates one. See docs/SKILLS.md in a Codewhale checkout.
/config is the live settings surface. Config filekeys are documented in docs/CONFIGURATION.md; provider/model routing in docs/PROVIDERS.md.
docs/KEYBINDINGS.md in a checkout. There is nokeybinding slash command; do not invent one.
codewhale-tui doctor reports the resolvedconfig path, provider credential presence (never values), and workspace state. Prefer its output over inference.
When the workspace is a Codewhale checkout, docs/ is present on disk and File with action: "read" is the right tool. Read the single most relevant file and quote the specific lines. Outside a checkout, docs/ is usually absent — in that case rely on /help, /config, and doctor, and say plainly that the reference docs are not installed locally.
docs/ looking for context.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-22 | pass→pass | 6,766 | 2,705 | -60% | 1 | 1 | 0% | 1,206 | 977 | -19% | 0 | 0 | — |
case-01 | fail→fail | 10,573 | 4,846 | -54% | 1 | 1 | 0% | 1,702 | 855 | -50% | 0 | 0 | — |
case-02 | fail→pass | 8,544 | 2,782 | -67% | 1 | 1 | 0% | 1,565 | 1,039 | -34% | 0 | 0 | — |
case-03 | fail→pass | 8,762 | 2,752 | -69% | 1 | 1 | 0% | 1,385 | 838 | -39% | 0 | 0 | — |
case-04 | fail→pass | 5,327 | 3,190 | -40% | 1 | 1 | 0% | 899 | 1,111 | +24% | 0 | 0 | — |
case-05 | fail→pass | 10,381 | 2,725 | -74% | 1 | 1 | 0% | 1,762 | 1,036 | -41% | 0 | 0 | — |
case-06 | fail→pass | 8,370 | 1,960 | -77% | 1 | 1 | 0% | 1,258 | 878 | -30% | 0 | 0 | — |
case-07 | fail→pass | 10,863 | 5,340 | -51% | 1 | 1 | 0% | 1,715 | 1,075 | -37% | 0 | 0 | — |
case-08 | pass→pass | 7,837 | 2,544 | -68% | 1 | 1 | 0% | 1,216 | 956 | -21% | 0 | 0 | — |
case-09 | fail→pass | 8,348 | 1,869 | -78% | 1 | 1 | 0% | 1,176 | 875 | -26% | 0 | 0 | — |
case-10 | fail→pass | 9,986 | 1,707 | -83% | 1 | 1 | 0% | 1,497 | 857 | -43% | 0 | 0 | — |
case-11 | fail→pass | 7,807 | 1,847 | -76% | 1 | 1 | 0% | 1,315 | 881 | -33% | 0 | 0 | — |
case-12 | pass→pass | 4,099 | 4,314 | +5% | 1 | 1 | 0% | 565 | 1,338 | +137% | 0 | 0 | — |
case-13 | fail→pass | 8,878 | 2,780 | -69% | 1 | 1 | 0% | 1,621 | 1,053 | -35% | 0 | 0 | — |
case-14 | pass→fail | 6,024 | 5,937 | -1% | 1 | 1 | 0% | 1,042 | 1,218 | +17% | 0 | 0 | — |
case-15 | pass→pass | 5,825 | 1,857 | -68% | 1 | 1 | 0% | 1,040 | 926 | -11% | 0 | 0 | — |
case-16 | pass→pass | 4,121 | 1,272 | -69% | 1 | 1 | 0% | 698 | 765 | +10% | 0 | 0 | — |
case-17 | pass→pass | 8,000 | 2,816 | -65% | 1 | 1 | 0% | 1,261 | 805 | -36% | 0 | 0 | — |
case-18 | fail→pass | 7,488 | 2,804 | -63% | 1 | 1 | 0% | 1,315 | 1,006 | -23% | 0 | 0 | — |
case-19 | fail→pass | 7,010 | 1,909 | -73% | 1 | 1 | 0% | 1,272 | 881 | -31% | 0 | 0 | — |
case-20 | pass→pass | 8,321 | 4,555 | -45% | 1 | 1 | 0% | 1,527 | 1,397 | -9% | 0 | 0 | — |
case-21 | pass→pass | 9,046 | 2,473 | -73% | 1 | 1 | 0% | 1,484 | 991 | -33% | 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 21 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 +50 percentage points is the difference between those two pass rates over the 21 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/7/2026 | +39% |
Other measured skills in the registry, with their headline benchmark lift.