deploy-model
Unified Azure OpenAI model deployment skill with intelligent intent-based routing. Handles quick preset deployments, fully customized deployments (version/SKU/capacity/RAI policy), and capacity discovery across regions and projects. USE FOR: deploy model, deploy gpt, create deployment, model deployment, deploy openai model, set up model, provision model, find capacity, check model availability, where can I deploy, best region for model, capacity analysis. DO NOT USE FOR: listing existing deployments (use foundry_models_deployments_list MCP tool), deleting deployments, agent creation (use agent/create), project creation (use project/create).
94
92%
Does it follow best practices?
Impact
96%
2.66xAverage score across 3 eval scenarios
Passed
No known issues
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 covers all key dimensions well. It provides specific capabilities, comprehensive natural trigger terms, explicit use/don't-use guidance, and clear boundaries with other skills. The DO NOT USE FOR section with alternative skill references is a particularly strong pattern for reducing routing conflicts.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: quick preset deployments, fully customized deployments with specific parameters (version/SKU/capacity/RAI policy), and capacity discovery across regions and projects. Very detailed about what it does. | 3 / 3 |
Completeness | Clearly answers both 'what' (handles preset deployments, customized deployments, capacity discovery) and 'when' (explicit USE FOR and DO NOT USE FOR clauses with specific trigger scenarios). The DO NOT USE FOR section adds extra clarity for routing. | 3 / 3 |
Trigger Term Quality | Excellent coverage of natural trigger terms users would say: 'deploy model', 'deploy gpt', 'create deployment', 'find capacity', 'check model availability', 'where can I deploy', 'best region for model'. These are natural phrases a user would actually type. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive with clear boundaries. The DO NOT USE FOR section explicitly differentiates from listing deployments, deleting deployments, agent creation, and project creation, naming the alternative tools/skills to use instead. This significantly reduces conflict risk. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured routing skill that effectively serves as an entry point for model deployment workflows. Its strengths are excellent progressive disclosure, clear validation checkpoints, and concrete CLI commands. The main weakness is moderate verbosity in the intent detection section, where the ASCII decision tree and multiple routing tables could be consolidated without losing clarity.
Suggestions
Consolidate the ASCII decision tree and the routing rules table into a single representation to reduce redundancy — the routing rules table alone captures all the information in the decision tree.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The routing tables and intent detection are useful but somewhat verbose. The decision tree ASCII art, multiple routing tables, and the project selection confirmation templates add bulk. Some of this (like explaining what 'custom' keywords mean) could be tightened, though most content is non-trivial routing logic Claude wouldn't inherently know. | 2 / 3 |
Actionability | Provides concrete CLI commands for validation (az cognitiveservices model list, az cognitiveservices usage list), specific JSON field paths to extract (.model.skus[].name), exact quota computation formulas (available = limit - currentValue), and clear confirmation UI templates. The routing rules are specific and executable. | 3 / 3 |
Workflow Clarity | Multi-step processes are clearly sequenced with explicit validation checkpoints: project confirmation before deployment, pre-deployment validation of both SKU support and quota availability, and multi-mode chaining (capacity → deploy) with user decision points. The warning about never deploying without confirmation adds a safety gate. | 3 / 3 |
Progressive Disclosure | Excellent progressive disclosure structure. The SKILL.md serves as a clear routing overview with a quick reference table, then delegates to three well-signaled sub-skills (preset, customize, capacity) that are one level deep. Cross-references to the quota skill are also cleanly handled with a brief note rather than inline duplication. | 3 / 3 |
Total | 11 / 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/github-copilot-for-azure
- Commit
a46a937
- 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.