MicrosoftKnown publishermicrosoft/MicroHack/03-Azure/01-04-AI/08_Predictive_Maintenance/.github/copilot-instructions.mdMicroHack / 08_Predictive_Maintenance
Copilot instructions220 starsChanged 7 days agoMITView on GitHub
This repo is used for a workshop/exercise. If such folders exist in the workspace, treat them as off-limits. This workshop must work reliably with smaller models (e.g., GPT-5-mini). Follow these constraints: Add brief comments explaining C#-specific idioms: - ?? and ??= operators (null coalescing) - Primary constructors - await using (like Python's async with)
Where files are headingNew Copilot instructions files have more rules about branches and pull requests, and fewer about code style.
Of the Copilot instructions files we read, the 2,349 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository11 other files it gives its agents.
- 03-Azure/01-04-AI/01_Agentic_SDLC/.github/copilot-instructions.mdCopilot instructions
- 03-Azure/01-04-AI/02_Inventory_Planning_Fabric/.github/copilot-instructions.mdCopilot instructions
- 03-Azure/01-04-AI/03_Inventory_Planning_Agentic/.github/copilot-instructions.mdCopilot instructions
- 03-Azure/01-04-AI/04_Agentic_Contract_Lifecycle_Management/.github/copilot-instructions.mdCopilot instructions
- 03-Azure/01-04-AI/11_Healthcare_Regulatory_Intelligence/.github/copilot-instructions.mdCopilot instructions
- dotnet-upgradeSkill
- java-upgradeSkill
- {{skill-name-kebab-case}}Skill
- skill-creatorSkill
- frontend-designSkill
- reconciliationSkill
The file
We split this copilot-instructions.md into the rules it gives an agent
11 rules. About 0 are common, 9 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈9
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 10
Use the Foundry Agents SDK pattern (Azure.AI.Projects + Microsoft.Agents.AI), not direct ChatCompletions.
- Distinctive: line 20
If such folders exist in the workspace, treat them as off-limits.
- Distinctive: line 23
Ask clarifying questions only when required.
- Distinctive: line 25
Prefer small, verifiable changes.
- Distinctive: line 32
Prefer producing a small number of complete files per request (1-3) over partial snippets
- Distinctive: line 33
Keep responses structured: What you will change, Files, How to run/verify
- Distinctive: line 34
Avoid optional extras unless explicitly requested (no extra abstractions, no extra models)
- Distinctive: line 35
When uncertain, state assumptions explicitly and proceed with a safe default
- Distinctive: line 38
Always use these exact NuGet package versions:
- Very specific: line 59
Use the hardcoded mappings from .github/agents/agentplanning.agent.md - implement as in-memory dictionaries in FaultMappingService.
- Distinctive: line 63
Add brief comments explaining C#-specific idioms:
≈ “Common” is an estimate: at least 1 in 5 of the 3,779 Copilot instructions files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Copilot custom instructions (workshop mode)
This repo is used for a workshop/exercise.
Key goal
Help the developer build the solution from context and instructions, not by copying an existing completed implementation.
Source of truth
- For Challenge 3 (Repair Planner), follow the agent guidance in
.github/agents/agentplanning.agent.md. - Use the Foundry Agents SDK pattern (
Azure.AI.Projects+Microsoft.Agents.AI), not direct ChatCompletions.
Do not use solution code as a template
Unless the user explicitly asks to reveal or compare with a solution, do not open, reference, or reuse code from any folder intended to contain completed answers, for example:
walkthrough/**/RepairPlanner/**walkthrough/****/*solution*/****/*Solution*/****/reference/**
If such folders exist in the workspace, treat them as off-limits.
How to proceed instead
- Ask clarifying questions only when required.
- Implement incrementally in the exercise project.
- Prefer small, verifiable changes; keep logging helpful.
Smaller-model friendly (CRITICAL)
This workshop must work reliably with smaller models (e.g., GPT-5-mini). Follow these constraints:
Output constraints
- Prefer producing a small number of complete files per request (1-3) over partial snippets
- Keep responses structured: What you will change, Files, How to run/verify
- Avoid optional extras unless explicitly requested (no extra abstractions, no extra models)
- When uncertain, state assumptions explicitly and proceed with a safe default
Pinned versions (Challenge 3)
Always use these exact NuGet package versions:
Azure.AI.Projects→2.1.0-beta.4Azure.Identity→1.21.0Microsoft.Agents.AI→1.20.0Microsoft.Agents.AI.Foundry→1.20.0-preview.260831.1Microsoft.Extensions.AI→10.9.0Microsoft.Extensions.AI.Abstractions→10.9.0Microsoft.Azure.Cosmos→3.63.0Microsoft.Extensions.DependencyInjection→10.0.12Microsoft.Extensions.Logging→10.0.12Microsoft.Extensions.Logging.Console→10.0.12Newtonsoft.Json→13.0.4
Environment variables (Challenge 3)
AZURE_AI_PROJECT_ENDPOINTMODEL_DEPLOYMENT_NAMECOSMOS_ENDPOINTCOSMOS_KEYCOSMOS_DATABASE_NAME
Fault→skills/parts mappings
Use the hardcoded mappings from .github/agents/agentplanning.agent.md - implement as in-memory dictionaries in FaultMappingService.
Code style for Python developers
Add brief comments explaining C#-specific idioms:
??and??=operators (null coalescing)- Primary constructors
await using(like Python'sasync with)
What NOT to do
- ❌ Don't use
Azure.AI.Inference/ChatCompletionsClient- use Foundry Agents SDK - ❌ Don't generate partial code snippets - generate complete files
- ❌ Don't add extra abstractions not in the spec
- ❌ Don't use different package versions than specified
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.