Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Run validation (lint, format, tests) for affected services based on changed files
.claude/skills/nudgebee-validate/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -48% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 118% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -29% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -24% | 0% |
Run validation for services affected by current changes. Optional argument: $ARGUMENTS (specific service name to validate, e.g., api-server).
If $ARGUMENTS is provided, validate only that service. Otherwise, detect affected services from changed files:
bash# Get changed files vs main (both staged and unstaged) git diff --name-only main...HEAD git diff --name-only git diff --name-only --cached
Map files to services and their validation commands:
| Path prefix | Service | Validation command | Working directory | |---|---|---|---| | api-server/services/ | api-server | make validate | api-server/services/ | | ticket-server/ | ticket-server | make validate | ticket-server/ | | collector-server/cloud-collector/ | cloud-collector | make validate | collector-server/cloud-collector/ | | collector-server/k8s-collector/relay-server/ | relay-server | make validate | collector-server/k8s-collector/relay-server/ | | collector-server/k8s-collector/app/ | k8s-collector-app | make lint && make test | collector-server/k8s-collector/app/ | | llm/code-analysis/ | code-analysis | make check | llm/code-analysis/ | | llm/llm-server/ | llm-server | make validate | llm/llm-server/ | | llm/rag-server/ | rag-server | make lint && make test | llm/rag-server/ | | llm/benchmark/ | benchmark | poetry run pytest | llm/benchmark/ | | ml-k8s-server/ | ml-k8s-server | make lint && make test | ml-k8s-server/ | | auto-pilot/ | auto-pilot | poetry run black --check . && poetry run flake8 . | auto-pilot/ | | auto-pilot/sidecar/ | auto-pilot-sidecar | poetry run black --check . && poetry run flake8 . | auto-pilot/sidecar/ | | notifications-server/ | notifications-server | poetry run black --check . && poetry run flake8 . | notifications-server/ | | app/ | frontend | npm run lint2 | app/ |
For each affected service, run its validation command from the correct working directory. Capture both stdout and stderr. Use a timeout of 5 minutes per service.
Run independent service validations in parallel where possible.
Output a summary table:
## Validation Results
| Service | Status | Details |
|---------|--------|---------|
| api-server | PASS/FAIL | {brief detail or error} |
| app | PASS/FAIL | {brief detail or error} |
{N} service(s) checked, {P} passed, {F} failedIf any service fails, show the relevant error output and suggest fixes.
If the user invokes with --fix in $ARGUMENTS, attempt to auto-fix issues:
make fmt then re-validatepoetry run black . then re-validatenpm run lint2:fix then re-validate| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-15 | fail→fail | 10,105 | 2,768 | -73% | 1 | 1 | 0% | 1,567 | 1,270 | -19% | 0 | 0 | — |
case-01 | fail→fail | 2,876 | 5,711 | +99% | 1 | 1 | 0% | 373 | 1,100 | +195% | 0 | 0 | — |
case-02 | fail→fail | 8,909 | 4,131 | -54% | 1 | 1 | 0% | 1,045 | 1,078 | +3% | 0 | 0 | — |
case-03 | fail→fail | 3,711 | 5,169 | +39% | 1 | 1 | 0% | 459 | 1,142 | +149% | 0 | 0 | — |
case-04 | fail→fail | 7,290 | 1,937 | -73% | 1 | 1 | 0% | 1,235 | 1,192 | -3% | 0 | 0 | — |
case-05 | fail→pass | 13,480 | 1,733 | -87% | 1 | 1 | 0% | 2,165 | 1,120 | -48% | 0 | 0 | — |
case-06 | fail→pass | 3,994 | 3,080 | -23% | 1 | 1 | 0% | 634 | 1,379 | +118% | 0 | 0 | — |
case-07 | fail→fail | 16,364 | 2,285 | -86% | 1 | 1 | 0% | 2,819 | 1,232 | -56% | 0 | 0 | — |
case-08 | fail→fail | 8,238 | 2,999 | -64% | 1 | 1 | 0% | 1,300 | 1,346 | +4% | 0 | 0 | — |
case-09 | fail→pass | 11,532 | 1,454 | -87% | 1 | 1 | 0% | 1,888 | 1,051 | -44% | 0 | 0 | — |
case-10 | fail→fail | 10,956 | 1,553 | -86% | 1 | 1 | 0% | 1,848 | 1,078 | -42% | 0 | 0 | — |
case-11 | fail→pass | 10,477 | 2,270 | -78% | 1 | 1 | 0% | 1,663 | 1,189 | -29% | 0 | 0 | — |
case-12 | pass→pass | 8,037 | 3,162 | -61% | 1 | 1 | 0% | 1,315 | 1,353 | +3% | 0 | 0 | — |
case-13 | fail→fail | 10,050 | 2,207 | -78% | 1 | 1 | 0% | 1,618 | 1,249 | -23% | 0 | 0 | — |
case-14 | fail→pass | 8,765 | 1,676 | -81% | 1 | 1 | 0% | 1,413 | 1,076 | -24% | 0 | 0 | — |
case-16 | fail→pass | 9,704 | 1,689 | -83% | 1 | 1 | 0% | 1,418 | 1,052 | -26% | 0 | 0 | — |
case-17 | fail→pass | 4,818 | 2,516 | -48% | 1 | 1 | 0% | 823 | 1,212 | +47% | 0 | 0 | — |
case-18 | fail→fail | 6,754 | 1,926 | -71% | 1 | 1 | 0% | 1,170 | 1,159 | -1% | 0 | 0 | — |
case-19 | fail→pass | 8,870 | 1,797 | -80% | 1 | 1 | 0% | 1,386 | 1,170 | -16% | 0 | 0 | — |
case-20 | pass→pass | 8,436 | 2,511 | -70% | 1 | 1 | 0% | 1,460 | 1,321 | -10% | 0 | 0 | — |
case-21 | pass→fail | 4,736 | 4,414 | -7% | 1 | 1 | 0% | 773 | 976 | +26% | 0 | 0 | — |
case-22 | pass→fail | 9,355 | 4,394 | -53% | 1 | 1 | 0% | 1,817 | 1,044 | -43% | 0 | 0 | — |
case-23 | pass→pass | 7,818 | 8,624 | +10% | 1 | 1 | 0% | 1,499 | 2,431 | +62% | 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. 23 cases were attempted, and 19 counted toward the lift figure. The other 4 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 +26 percentage points is the difference between those two pass rates over the 19 comparable cases. 2 cases got worse with the skill loaded, and they are 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.