Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Manage environment variables securely. Handles distinction between .env (template) and .env.local (secrets).
.claude/skills/aiskillstore-env-handler/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 82% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -26% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -18% | 0% |
.env.example: Do not create this file. Use .env as the template..env.local: Actual sensitive values must live in .env.local (git-ignored)..env.local MUST have a corresponding entry in .env.KEY=""KEY="default_value"When you need to add a secret (e.g., REPLICATE_API_TOKEN):
.env:Add the variable with an empty string value. bash # .env REPLICATE_API_TOKEN=""
Explicitly request the user to add the actual value to their local secrets file. > "I have added REPLICATE_API_TOKEN to your .env file. Please open .env.local and add the actual token: REPLICATE_API_TOKEN=your_token_here"
When adding a public or configuration variable (e.g., NEXT_PUBLIC_APP_URL):
.env:Add the variable with its default or development value. bash # .env NEXT_PUBLIC_APP_URL="http://localhost:3000"
process.env.KEYprocess.env.NEXT_PUBLIC_KEY.env?.env empty?.env.local?| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 7,204 | 2,349 | -67% | 1 | 1 | 0% | 1,271 | 842 | -34% | 0 | 0 | — |
case-11 | pass→pass | 2,831 | 2,441 | -14% | 1 | 1 | 0% | 421 | 806 | +91% | 0 | 0 | — |
case-02 | fail→pass | 3,007 | 2,632 | -12% | 1 | 1 | 0% | 483 | 879 | +82% | 0 | 0 | — |
case-03 | fail→pass | 7,680 | 2,972 | -61% | 1 | 1 | 0% | 1,308 | 812 | -38% | 0 | 0 | — |
case-04 | fail→pass | 7,619 | 3,092 | -59% | 1 | 1 | 0% | 1,373 | 1,018 | -26% | 0 | 0 | — |
case-05 | pass→pass | 9,920 | 3,902 | -61% | 1 | 1 | 0% | 1,898 | 1,178 | -38% | 0 | 0 | — |
case-10 | fail→pass | 6,273 | 2,180 | -65% | 1 | 1 | 0% | 1,015 | 834 | -18% | 0 | 0 | — |
case-06 | pass→pass | 6,530 | 5,241 | -20% | 1 | 1 | 0% | 1,253 | 1,430 | +14% | 0 | 0 | — |
case-07 | pass→pass | 8,273 | 3,015 | -64% | 1 | 1 | 0% | 1,618 | 965 | -40% | 0 | 0 | — |
case-08 | fail→pass | 11,060 | 4,128 | -63% | 1 | 1 | 0% | 1,675 | 1,178 | -30% | 0 | 0 | — |
case-09 | fail→pass | 9,712 | 4,057 | -58% | 1 | 1 | 0% | 1,628 | 1,154 | -29% | 0 | 0 | — |
case-12 | fail→pass | 9,784 | 2,838 | -71% | 1 | 1 | 0% | 1,730 | 894 | -48% | 0 | 0 | — |
case-13 | fail→pass | 2,341 | 3,434 | +47% | 1 | 1 | 0% | 367 | 1,034 | +182% | 0 | 0 | — |
case-14 | pass→pass | 7,750 | 3,123 | -60% | 1 | 1 | 0% | 1,326 | 939 | -29% | 0 | 0 | — |
case-15 | fail→pass | 8,001 | 4,750 | -41% | 1 | 1 | 0% | 1,378 | 1,272 | -8% | 0 | 0 | — |
case-16 | pass→pass | 8,244 | 3,788 | -54% | 1 | 1 | 0% | 1,546 | 1,022 | -34% | 0 | 0 | — |
case-17 | pass→pass | 5,025 | 3,366 | -33% | 1 | 1 | 0% | 866 | 995 | +15% | 0 | 0 | — |
case-18 | fail→pass | 7,268 | 2,103 | -71% | 1 | 1 | 0% | 1,352 | 775 | -43% | 0 | 0 | — |
case-19 | pass→pass | 9,008 | 2,126 | -76% | 1 | 1 | 0% | 1,552 | 788 | -49% | 0 | 0 | — |
case-20 | pass→pass | 13,942 | 10,231 | -27% | 1 | 1 | 0% | 2,443 | 2,269 | -7% | 0 | 0 | — |
case-21 | pass→pass | 7,741 | 8,195 | +6% | 1 | 1 | 0% | 1,409 | 1,934 | +37% | 0 | 0 | — |
case-22 | pass→fail | 8,998 | 8,702 | -3% | 1 | 1 | 0% | 1,778 | 2,165 | +22% | 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 +45 percentage points is the difference between those two pass rates over the 22 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.
Other measured skills in the registry, with their headline benchmark lift.