AI Sidekick
AI-powered asset analysis and explanation using cloud-hosted AI infrastructure with enhanced context from centralized metadata.
Overview
AI Sidekick provides intelligent analysis of assets using:
- Project metadata from cloud storage
- Glossary terms and definitions
- Lineage information across projects
- Test results and history
- Historical patterns and best practices
Explaining an Asset
Endpoint
POST /api/msh/ai/explain
Generate a natural language explanation of what an asset does.
Request:
curl -X POST \
-H "Authorization: Bearer $API_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"asset_id": 5,
"include_tests": true,
"include_history": true
}' \
https://api.msh.io/api/msh/ai/explain
Response:
{
"success": true,
"asset_id": 5,
"summary": "This asset calculates monthly recurring revenue by aggregating active subscription amounts...",
"grain": "month",
"upstream_assets": ["subscriptions", "plans"],
"downstream_assets": ["revenue_dashboard"],
"business_terms": ["term.customer", "metric.mrr"],
"policies": ["policy.pii_encryption"]
}
Reviewing an Asset
Endpoint
POST /api/msh/ai/review
Review an asset for risks, issues, and best practice violations.
Request:
curl -X POST \
-H "Authorization: Bearer $API_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"asset_id": 5
}' \
https://api.msh.io/api/msh/ai/review
Response:
{
"success": true,
"asset_id": 5,
"summary": "Asset is well-structured with proper tests...",
"risks": [
"No PII encryption policy applied",
"Missing glossary link for 'customer_id' column"
],
"glossary_issues": [
"Column 'customer_id' should be linked to term.customer"
],
"performance_notes": [
"Consider adding index on 'subscription_date' column"
],
"suggested_changes": [
{
"type": "add_test",
"description": "Add test for revenue > 0"
}
]
}
Analyzing Lineage Impact
Endpoint
POST /api/msh/ai/lineage-impact
Analyze downstream impact of changes to an asset.
Request:
curl -X POST \
-H "Authorization: Bearer $API_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"asset_id": 5
}' \
https://api.msh.io/api/msh/ai/lineage-impact
Response:
{
"success": true,
"asset_id": 5,
"downstream_impact": {
"direct_dependents": 3,
"transitive_dependents": 12,
"affected_assets": ["revenue_dashboard", "forecast_model"]
},
"breaking_changes": [],
"warnings": [
"Schema change detected in 'amount' column type"
],
"recommendations": [
"Run tests on downstream assets before deploying",
"Notify owners of affected assets"
]
}
Use Cases
Pre-Deployment Review
Review assets before deploying changes:
# Review asset
curl -X POST \
-H "Authorization: Bearer $API_TOKEN" \
-H "Content-Type: application/json" \
-d '{"asset_id": 5}' \
https://api.msh.io/api/msh/ai/review
# Analyze impact
curl -X POST \
-H "Authorization: Bearer $API_TOKEN" \
-H "Content-Type: application/json" \
-d '{"asset_id": 5}' \
https://api.msh.io/api/msh/ai/lineage-impact
Understanding Legacy Assets
Explain what existing assets do:
curl -X POST \
-H "Authorization: Bearer $API_TOKEN" \
-H "Content-Type: application/json" \
-d '{"asset_id": 5, "include_tests": true}' \
https://api.msh.io/api/msh/ai/explain
Code Quality Assurance
Review assets for best practices:
curl -X POST \
-H "Authorization: Bearer $API_TOKEN" \
-H "Content-Type: application/json" \
-d '{"asset_id": 5}' \
https://api.msh.io/api/msh/ai/review
Differences from OSS
The Cloud AI Sidekick provides:
- Enhanced Context: Uses centralized metadata and glossary
- No API Keys: Uses cloud-hosted AI infrastructure
- Cross-Project Analysis: Analyzes dependencies across projects
- Historical Patterns: Learns from historical data
Related Documentation
- AI Features - OSS AI features
- Semantic Search - Natural language search
- Pre-Deploy Safety - Safety checks