Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Fetch up-to-date library documentation via Context7 API. Use PROACTIVELY when: (1) Working with ANY external library (React, Next.js, Supabase, etc.) (2) User asks about library APIs, patterns, or best practices (3) Implementing features that rely on third-party packages (4) Debugging library-specific issues (5) Need current documentation beyond training data cutoff Always prefer this over guessing library APIs or using outdated knowledge.
.claude/skills/sundial-org-context7-2/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -31% | 0% |
Retrieve current library documentation via Context7 API.
bashpython3 ~/.claude/skills/context7/scripts/context7.py search "<library-name>"
Example:
bashpython3 ~/.claude/skills/context7/scripts/context7.py search "next.js"
Returns library metadata including the id field needed for step 2.
bashpython3 ~/.claude/skills/context7/scripts/context7.py context "<library-id>" "<query>"
Example:
bashpython3 ~/.claude/skills/context7/scripts/context7.py context "/vercel/next.js" "app router middleware"
Options:
--type txt|md - Output format (default: txt)--tokens N - Limit response tokens| Task | Command | |------|---------| | Find React docs | search "react" | | Get React hooks info | context "/facebook/react" "useEffect cleanup" | | Find Supabase | search "supabase" | | Get Supabase auth | context "/supabase/supabase" "authentication row level security" |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→fail | 12,325 | 32,524 | +164% | 1 | 1 | 0% | 2,048 | 5,645 | +176% | 0 | 0 | — |
case-16 | fail→fail | 8,025 | 5,284 | -34% | 1 | 1 | 0% | 1,451 | 596 | -59% | 0 | 0 | — |
case-01 | fail→fail | 14,576 | 5,189 | -64% | 1 | 1 | 0% | 2,436 | 521 | -79% | 0 | 0 | — |
case-03 | fail→fail | 13,529 | 4,671 | -65% | 1 | 1 | 0% | 2,165 | 514 | -76% | 0 | 0 | — |
case-04 | fail→pass | 3,647 | 1,816 | -50% | 1 | 1 | 0% | 523 | 588 | +12% | 0 | 0 | — |
case-05 | fail→pass | 6,684 | 2,384 | -64% | 1 | 1 | 0% | 1,007 | 661 | -34% | 0 | 0 | — |
case-06 | fail→pass | 11,310 | 2,615 | -77% | 1 | 1 | 0% | 1,619 | 664 | -59% | 0 | 0 | — |
case-22 | pass→fail | 3,677 | 2,743 | -25% | 1 | 1 | 0% | 511 | 492 | -4% | 0 | 0 | — |
case-07 | fail→pass | 7,276 | 2,340 | -68% | 1 | 1 | 0% | 1,040 | 657 | -37% | 0 | 0 | — |
case-08 | pass→pass | 3,479 | 2,803 | -19% | 1 | 1 | 0% | 465 | 722 | +55% | 0 | 0 | — |
case-09 | fail→fail | 11,035 | 4,954 | -55% | 1 | 1 | 0% | 1,745 | 485 | -72% | 0 | 0 | — |
case-10 | fail→fail | 12,280 | 4,945 | -60% | 1 | 1 | 0% | 2,076 | 538 | -74% | 0 | 0 | — |
case-11 | fail→fail | 10,269 | 4,586 | -55% | 1 | 1 | 0% | 1,533 | 511 | -67% | 0 | 0 | — |
case-12 | fail→fail | 11,519 | 4,861 | -58% | 1 | 1 | 0% | 1,203 | 530 | -56% | 0 | 0 | — |
case-13 | fail→pass | 8,625 | 4,494 | -48% | 1 | 1 | 0% | 1,316 | 907 | -31% | 0 | 0 | — |
case-14 | pass→pass | 5,152 | 2,460 | -52% | 1 | 1 | 0% | 673 | 703 | +4% | 0 | 0 | — |
case-15 | fail→fail | 26,649 | 4,535 | -83% | 1 | 1 | 0% | 3,393 | 508 | -85% | 0 | 0 | — |
case-17 | fail→pass | 5,037 | 2,052 | -59% | 1 | 1 | 0% | 637 | 640 | +0% | 0 | 0 | — |
case-18 | fail→pass | 10,225 | 1,695 | -83% | 1 | 1 | 0% | 1,441 | 548 | -62% | 0 | 0 | — |
case-19 | fail→fail | 7,716 | 4,950 | -36% | 1 | 1 | 0% | 517 | 527 | +2% | 0 | 0 | — |
case-20 | pass→pass | 6,409 | 4,799 | -25% | 1 | 1 | 0% | 1,014 | 998 | -2% | 0 | 0 | — |
case-21 | fail→fail | 3,107 | 6,349 | +104% | 1 | 1 | 0% | 286 | 741 | +159% | 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 12 counted toward the lift figure. The other 10 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 +27 percentage points is the difference between those two pass rates over the 12 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 | 8/19/2026 | +23% |
Other measured skills in the registry, with their headline benchmark lift.