saga-architecture
Skill0 starsChanged 24 days agoMITView on GitHub
Saga coordination patterns for distributed transactions with compensating actions.
In this repository32 other files it gives its agents.
- .claude/CLAUDE.mdCLAUDE.md
- api-designSkill
- arch-constraint-guardSkill
- backend-patternsSkill
- clean-architectureSkill
- codebase-mappingSkill
- context-budgetSkill
- cqrsSkill
- ddd-architectureSkill
- debug-flowSkill
- dependency-auditSkill
- django-patternsSkill
- django-tddSkill
- event-driven-architectureSkill
- fd-codenavSkill
- fd-mapSkill
- frontend-patternSkill
- git-releaseSkill
- git-workflowSkill
- golang-patternsSkill
- hexagonal-architectureSkill
- java-patternsSkill
- layered-architectureSkill
- performance-profilingSkill
- postgres-patternsSkill
- python-patternsSkill
- refactor-guideSkill
- research-firstSkill
- rust-patternsSkill
- security-scanSkill
- tdd-workflowSkill
- test-gap-detectorSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 10
When coordinating distributed operations across multiple services or data stores where ACID transactions are not available and compensating …
- Distinctive: line 16
Implement idempotent operations: - Ensure each step can be safely retried and compensation can be safely reapplied.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill 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
saga-architecture
When to Activate
When coordinating distributed operations across multiple services or data stores where ACID transactions are not available and compensating actions are needed to maintain eventual consistency.
Steps
- Identify the saga participants - Determine which services or components participate in the distributed operation.
- Define the saga choreography or orchestration - Choose whether sagas will be choreographed (event-driven) or orchestrated (central coordinator).
- Define each step with corresponding compensation - For every forward action, define what compensating action undoes it.
- Implement idempotent operations - Ensure each step can be safely retried and compensation can be safely reapplied.
- Handle saga failures with compensation - On failure, execute compensations in reverse order (for orchestrating sagas) or react to failure events (for choreographing sagas).
- Persist saga state - Store saga state to survive process crashes and enable recovery.
- Add timeout and retry logic - Detect stuck sagas and advance or compensate accordingly.
Examples
// Saga State
interface SagaState<T> {
id: string
currentStep: number
data: T
status: 'pending' | 'in_progress' | 'completed' | 'compensating' | 'failed'
}
// Orchestrating Saga - Central coordinator manages steps
class OrderProcessingSaga {
private readonly steps: SagaStep[]
constructor(
private readonly sagaOrchestrator: SagaOrchestrator,
private readonly inventoryService: InventoryService,
private readonly paymentService: PaymentService,
private readonly shippingService: ShippingService
) {
this.steps = [
{
name: 'reserve_inventory',
execute: (state) => this.inventoryService.reserve(state.orderId, state.items),
compensate: (state) => this.inventoryService.release(state.orderId, state.items)
},
{
name: 'process_payment',
execute: (state) => this.paymentService.charge(state.orderId, state.total),
compensate: (state) => this.paymentService.refund(state.orderId, state.total)
},
{
name: 'initiate_shipping',
execute: (state) => this.shippingService.createShipment(state.orderId),
compensate: (state) => this.shippingService.cancelShipment(state.shipmentId)
}
]
}
async execute(orderId: string): Promise<void> {
const state: SagaState<OrderSagaData> = {
id: generateId(),
currentStep: 0,
data: { orderId, items: [], total: 0 },
status: 'in_progress'
}
await this.sagaOrchestrator.start(state, this.steps)
}
}
// Choreography-based Saga - Events trigger reactions
class OrderCreatedHandler {
constructor(private readonly eventBus: EventBus) {}
async handle(event: OrderCreatedEvent): Promise<void> {
// Step 1: Reserve inventory
try {
await this.inventoryService.reserve(event.orderId, event.items)
this.eventBus.publish(new InventoryReservedEvent(event.orderId))
} catch (error) {
this.eventBus.publish(new InventoryReservationFailedEvent(event.orderId, error.message))
}
}
}
class InventoryReservedHandler {
async handle(event: InventoryReservedEvent): Promise<void> {
// Step 2: Process payment
try {
await this.paymentService.charge(event.orderId, event.total)
this.eventBus.publish(new PaymentProcessedEvent(event.orderId))
} catch (error) {
// Compensate by releasing inventory
this.eventBus.publish(new InventoryReleaseRequestedEvent(event.orderId))
}
}
}
// Idempotent Step Implementation
class PaymentService {
async charge(orderId: string, amount: Money): Promise<TransactionId> {
const existingTx = await this.transactionRepo.findByOrderId(orderId)
if (existingTx) {
return existingTx.id // Idempotent: return existing instead of charging again
}
const transaction = await this.paymentGateway.charge(amount)
await this.transactionRepo.save({ orderId, transaction })
return transaction.id
}
}Related Skills
- clean-architecture
- hexagonal-architecture
- ddd-architecture
- backend-patterns
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.

