Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Expert knowledge for Azure Anyscale On Azure development including limits & quotas, security, and configuration. Use when setting up Anyscale image builds, VNET/subnet networking, Azure AD auth, RBAC roles, or region availability, and other Azure Anyscale On Azure related development tasks. Not for Azure Kubernetes Service (AKS) (use azure-kubernetes-service), Azure Container Apps (use azure-container-apps), Azure Batch (use azure-batch), Azure Virtual Machines (use azure-virtual-machines).
.claude/skills/microsoftdocs-azure-anyscale-on-azure/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -42% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -53% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 4% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 123% | 0% |
This skill provides expert guidance for Azure Anyscale On Azure. Covers limits & quotas, security, and configuration. It combines local quick-reference content with remote documentation fetching capabilities.
> IMPORTANT for Agent: Use the Category Index below to locate relevant sections. For categories with line ranges (e.g., L35-L120), use read_file with the specified lines. For categories with file links (e.g., [security.md](security.md)), use read_file on the linked reference file
> IMPORTANT for Agent: If metadata.generated_at is more than 3 months old, suggest the user pull the latest version from the repository. If mcp_microsoftdocs tools are not available, suggest the user install it: Installation Guide
This skill requires network access to fetch documentation content:
mcp_microsoftdocs:microsoft_docs_fetch with query string from=learn-agent-skill. Returns Markdown.fetch_webpage with query string from=learn-agent-skill&accept=text/markdown. Returns Markdown.| Category | Lines | Description | |----------|-------|-------------| | Limits & Quotas | L31-L35 | Supported Azure regions for deploying and running Anyscale on Azure, including how to check regional availability and constraints. | | Security | L36-L40 | Configuring identity, Azure AD integration, and role-based access control (RBAC) for secure access and permissions management in Anyscale on Azure. | | Configuration | L41-L45 | Configuring Anyscale on Azure clouds: enabling container image builds and setting up required networking (VNETs, subnets, connectivity, security) for deployments. |
| Topic | URL | |-------|-----| | Check supported Azure regions for Anyscale | https://learn.microsoft.com/en-us/azure/anyscale-on-azure/supported-regions |
| Topic | URL | |-------|-----| | Set up identity and RBAC for Anyscale on Azure | https://learn.microsoft.com/en-us/azure/anyscale-on-azure/identity-access |
| Topic | URL | |-------|-----| | Enable container image builds for Anyscale clouds | https://learn.microsoft.com/en-us/azure/anyscale-on-azure/configure-container-image-builds | | Configure networking for Anyscale on Azure deployments | https://learn.microsoft.com/en-us/azure/anyscale-on-azure/networking |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 29,502 | 35,037 | +19% | 1 | 1 | 0% | 4,311 | 4,443 | +3% | 0 | 0 | — |
case-02 | fail→fail | 15,828 | 15,233 | -4% | 1 | 1 | 0% | 2,196 | 1,145 | -48% | 0 | 0 | — |
case-03 | fail→fail | 24,458 | 26,028 | +6% | 1 | 1 | 0% | 2,733 | 3,072 | +12% | 0 | 0 | — |
case-04 | fail→fail | 21,023 | 18,597 | -12% | 1 | 1 | 0% | 3,001 | 1,106 | -63% | 0 | 0 | — |
case-05 | fail→fail | 16,331 | 17,312 | +6% | 1 | 1 | 0% | 1,962 | 966 | -51% | 0 | 0 | — |
case-06 | fail→fail | 31,585 | 13,634 | -57% | 1 | 1 | 0% | 1,978 | 1,204 | -39% | 0 | 0 | — |
case-07 | fail→pass | 18,380 | 8,427 | -54% | 1 | 1 | 0% | 1,766 | 1,027 | -42% | 0 | 0 | — |
case-08 | fail→pass | 19,621 | 2,849 | -85% | 1 | 1 | 0% | 2,509 | 1,173 | -53% | 0 | 0 | — |
case-09 | fail→pass | 14,215 | 2,078 | -85% | 1 | 1 | 0% | 1,312 | 1,015 | -23% | 0 | 0 | — |
case-10 | fail→pass | 12,439 | 7,975 | -36% | 1 | 1 | 0% | 1,102 | 1,150 | +4% | 0 | 0 | — |
case-11 | fail→pass | 8,106 | 8,474 | +5% | 1 | 1 | 0% | 461 | 1,028 | +123% | 0 | 0 | — |
case-12 | fail→pass | 16,788 | 8,244 | -51% | 1 | 1 | 0% | 2,285 | 1,456 | -36% | 0 | 0 | — |
case-13 | fail→pass | 10,145 | 6,875 | -32% | 1 | 1 | 0% | 1,295 | 964 | -26% | 0 | 0 | — |
case-14 | fail→pass | 26,107 | 7,874 | -70% | 1 | 1 | 0% | 2,993 | 989 | -67% | 0 | 0 | — |
case-15 | fail→pass | 7,160 | 1,817 | -75% | 1 | 1 | 0% | 1,156 | 938 | -19% | 0 | 0 | — |
case-16 | fail→pass | 12,036 | 2,079 | -83% | 1 | 1 | 0% | 1,833 | 982 | -46% | 0 | 0 | — |
case-17 | fail→fail | 24,808 | 36,233 | +46% | 1 | 1 | 0% | 3,488 | 3,195 | -8% | 0 | 0 | — |
case-18 | fail→pass | 11,046 | 8,359 | -24% | 1 | 1 | 0% | 2,099 | 1,055 | -50% | 0 | 0 | — |
case-19 | fail→pass | 12,932 | 3,156 | -76% | 1 | 1 | 0% | 2,360 | 1,266 | -46% | 0 | 0 | — |
case-20 | pass→pass | 17,216 | 32,593 | +89% | 1 | 1 | 0% | 3,283 | 4,644 | +41% | 0 | 0 | — |
case-21 | pass→pass | 18,723 | 10,217 | -45% | 1 | 1 | 0% | 2,466 | 2,568 | +4% | 0 | 0 | — |
case-22 | pass→pass | 13,404 | 13,112 | -2% | 1 | 1 | 0% | 2,478 | 3,093 | +25% | 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 18 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 +55 percentage points is the difference between those two pass rates over the 18 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.