Prompts Reference¶
MCP prompts for common Contree workflows. Prompts provide structured instructions that guide AI agents through multi-step tasks.
Quick Reference¶
Prompt |
Description |
Key Parameters |
|---|---|---|
Prepare container environment with CHECK-PREPARE-EXECUTE flow |
|
|
Run Python code in isolated container |
|
|
Run shell command in isolated container |
|
|
Sync local files and run command |
|
|
Install packages and create reusable image |
|
|
Run multiple tasks in parallel |
|
|
Build project: install deps and run tests |
|
|
Diagnose failed operation |
|
|
Explore container image contents |
|
|
Multi-stage build with rollback points |
|
Using Prompts¶
With MCP Clients¶
MCP-compatible clients can invoke prompts directly:
{
"prompt": "prepare-environment",
"args": {
"task": "Train ML model",
"base": "python:3.11-slim",
"packages": "numpy pandas scikit-learn"
}
}
Prompt Output¶
Prompts return structured instructions that guide the agent through:
Step-by-step workflows - Ordered operations with clear dependencies
Tool selection - Which Contree tools to use and when
Parameter guidance - Correct values for each tool call
Best practices - Following the CHECK-PREPARE-EXECUTE pattern
Categories¶
Environment Setup¶
prepare-environment - Full workflow with environment reuse
install-packages - Install and tag for reuse
Code Execution¶
run-python - Quick Python execution
run-shell - Shell command execution
sync-and-run - Local files + execution
Building and Testing¶
build-project - Standard build + test workflow
multi-stage-build - Complex builds with checkpoints
Operations¶
parallel-tasks - Concurrent execution
debug-failure - Error diagnosis
inspect-image - Image exploration