azure-hosted-copilot-sdk
Build, deploy, and modify GitHub Copilot SDK apps on Azure. MANDATORY when codebase contains @github/copilot-sdk or CopilotClient in package.json. PREFER OVER azure-prepare when copilot-sdk markers detected. WHEN: copilot SDK, @github/copilot-sdk, copilot-powered app, build copilot app, prepare copilot app, add feature to copilot app, modify copilot app, BYOM, bring your own model, CopilotClient, createSession, sendAndWait, azd init copilot. DO NOT USE FOR: deploying already-prepared copilot-sdk apps (use azure-deploy), general web apps without copilot SDK (use azure-prepare), Copilot Extensions, Foundry agents (use microsoft-foundry).
68
81%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Advisory
Suggest reviewing before use
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is an excellent skill description that excels across all dimensions. It provides specific actions, comprehensive trigger terms, explicit when/when-not guidance, and clear differentiation from related skills. The DO NOT USE FOR and PREFER OVER clauses are particularly effective at reducing skill selection conflicts.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: 'Build, deploy, and modify GitHub Copilot SDK apps on Azure.' Also specifies concrete markers like '@github/copilot-sdk', 'CopilotClient', 'createSession', 'sendAndWait', and 'azd init copilot'. | 3 / 3 |
Completeness | Clearly answers both 'what' (build, deploy, modify GitHub Copilot SDK apps on Azure) and 'when' (explicit WHEN clause with trigger terms, MANDATORY conditions, and a DO NOT USE FOR section that further clarifies boundaries). This is exceptionally thorough. | 3 / 3 |
Trigger Term Quality | Excellent coverage of natural trigger terms users would say, including package names ('@github/copilot-sdk', 'CopilotClient'), action phrases ('build copilot app', 'modify copilot app'), and acronyms ('BYOM', 'bring your own model'). The WHEN clause explicitly enumerates many realistic trigger terms. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive with explicit boundary-setting: the DO NOT USE FOR section names specific alternative skills (azure-deploy, azure-prepare, microsoft-foundry) and clarifies exactly when this skill should NOT be selected. The PREFER OVER clause further reduces conflict risk. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
62%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill provides a well-structured orchestration guide for GitHub Copilot SDK apps on Azure with clear routing logic and workflow sequencing. Its main weakness is verbosity in the detection/routing section, which repeats the same priority rules multiple times with warning callouts, and the lack of inline executable code examples — most actionable details are deferred to reference files that weren't provided for verification.
Suggestions
Consolidate the three repeated warning callouts about routing priority into a single concise statement at the top of the detection section
Add at least one inline executable code snippet for the most common path (e.g., BYOM provider config with bearerToken) rather than deferring all code to references
Remove the 'Generic prompts that MUST trigger this skill' table — this is routing metadata that belongs in frontmatter or a separate routing config, not in the skill body
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The detection section is quite verbose with repeated warnings and redundant tables explaining routing logic. The scaffolding and deployment steps are lean and efficient, but the mandatory first check section could be significantly tightened — the same routing priority is stated three different ways with warning callouts. | 2 / 3 |
Actionability | Provides concrete commands (e.g., `azd init --template`), specific detection markers, and clear routing tables. However, most steps delegate to reference files for actual implementation details, and there are no executable code examples for SDK usage patterns (e.g., BYOM config code is deferred to references). The skill tells you what to do at a high level but lacks copy-paste-ready code. | 2 / 3 |
Workflow Clarity | The workflow is clearly sequenced: detect → route → scaffold → configure model → deploy. The routing table in Step 1 covers all user intent scenarios, Step 4 explicitly sequences sub-skill invocation (azure-prepare → azure-validate → azure-deploy), and there are clear yield conditions (e.g., yield to azure-deploy when deployment-plan.md exists). The detection procedure has explicit ordered steps with fallback scanning. | 3 / 3 |
Progressive Disclosure | References to supporting files (references/copilot-sdk.md, references/deploy-existing.md, references/azure-model-config.md, references/auth-best-practices.md, references/existing-project-integration.md) are well-signaled and one level deep. However, no bundle files were provided, so we cannot verify these references exist. The main body keeps a good overview structure, but the detection section is overly detailed inline content that could potentially be extracted. | 2 / 3 |
Total | 9 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- microsoft/azure-skills
- Commit
d02fd24
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.