resonate-saga-pattern-go

resonate-saga-pattern-go is a skill for Claude Code, Codex from resonatehq/resonate-skills. It costs 86 tokens per session (2,832 once invoked), scanned A, original, Apache-2.0.

A Go implementation of the saga pattern for Resonate workflows. It divides a distributed transaction into steps and runs each completed step's undo action in reverse order if a later step fails.

In plain words
What is it for?
Use it for multi-step Go workflows where intermediate results are visible to other systems, such as ordering or payment flows. It shows how to track completed steps and dispatch their compensations.
Why use it?
It helps maintain consistency when work crosses services and cannot use one shared database transaction. The workflow can recover by compensating for actions that already happened.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it for multi-step Go workflows where intermediate results are visible to other systems, such as ordering or payment flows. It shows how to track completed steps and dispatch their compensations.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/resonatehq/resonate-skills/resonate-saga-pattern-go
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add resonatehq/resonate-skills --skill resonate-saga-pattern-go
Clone the repo
git clone --depth 1 https://github.com/resonatehq/resonate-skills

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for resonate-saga-pattern-go

README.md
[![agentmods](https://agentmods.dev/badge/skills/resonatehq/resonate-skills/resonate-saga-pattern-go/github.svg)](https://agentmods.dev/skills/resonatehq/resonate-skills/resonate-saga-pattern-go)
Your own site
<a href="https://agentmods.dev/skills/resonatehq/resonate-skills/resonate-saga-pattern-go"><img src="https://agentmods.dev/badge/skills/resonatehq/resonate-skills/resonate-saga-pattern-go/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for resonate-saga-pattern-go

Your own site · 80×15
<a href="https://agentmods.dev/skills/resonatehq/resonate-skills/resonate-saga-pattern-go"><img src="https://agentmods.dev/badge/skills/resonatehq/resonate-skills/resonate-saga-pattern-go.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 86 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,832 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00086 $0.02832
Opus 5.5 $0.00034 $0.01133
Sonnet 5.5 $0.00017 $0.00566
Haiku 4.5 $0.00009 $0.00283

Measured 2mo ago against content hash a226f50b47b6, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

resonate-saga-pattern-go scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

resonate-saga-pattern-go/SKILL.md · 288 lines

How it starts

The opening of the file, as written. The whole thing — 288 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Resonate Saga Pattern — Go

Version note. The Go SDK's first tagged release is 0.1.0. The tag is 0.1.0, not v0.1.0, so go get …@latest does not resolve to it — pin the tag explicitly: go get github.com/resonatehq/[email protected]. APIs may still change before a 1.0. There is no Go saga example repo yet, so all snippets use only documented SDK surface verified against the 0.1.0 tag source.

Overview

A saga is a long-running transaction split into discrete steps, each with a compensating action. Forward steps execute top-to-bottom; on any failure, completed compensations run bottom-to-top. Because the whole workflow body re-runs on resume, the completed []Step slice is reconstructed deterministically on replay — each ctx.Run call short-circuits its settled durable promise in the same order, so the slice is rebuilt identically before reaching the point of failure.

For the language-agnostic mental model — including choreography vs orchestration, idempotency requirements, and when not to use sagas — see resonate-saga-pattern-typescript.

When to use

  • Multi-step workflow where intermediate state is visible to other systems between steps
  • Each step is individually compensable (inventory hold, payment charge, shipment create)
  • You need "all or nothing" consistency without a distributed 2PC transaction
  • Steps span services and may take seconds to minutes each
  • Compensation logic exists and is idempotent for every committed step

Basic shape

package main

import (
    "context"
    "errors"
    "fmt"
    "log"
    "time"

    resonate "github.com/resonatehq/resonate-sdk-go"
)

// Step is a string enum for the forward steps that have been committed.
// Using type Step string + a switch gives idiomatic Go compensation dispatch.
type Step string

const (
    StepInventory Step = "inventory"
    StepPayment   Step = "payment"
    StepShipment  Step = "shipment"
)

type OrderArgs struct {
    OrderID string `json:"order_id"`
}

type SagaResult struct {
    Status      string   `json:"status"`
    OrderID     string   `json:"order_id"`
    Compensated []string `json:"compensated,omitempty"`
}

type CompensateArgs struct {
    Step    Step   `json:"step"`
    OrderID string `json:"order_id"`
}

// placeOrder is the saga orchestrator. It runs forward steps in sequence,
// tracking each committed step. On the first failure it runs compensations in
// reverse, each as its own durable ctx.Run so they are retryable.
func placeOrder(ctx *resonate.Context, args OrderArgs) (SagaResult, error) {
    var completed []Step

    // --- forward path ---
    fInv, err := ctx.Run(reserveInventory, args.OrderID)
    if err != nil {
        return SagaResult{}, err
    }
    if err := fInv.Await(nil); err != nil {
        return compensateAll(ctx, args.OrderID, completed, err)
    }
    completed = append(completed, StepInventory)

    fPay, err := ctx.Run(chargePayment, args.OrderID)
    if err != nil {
        return SagaResult{}, err
    }
    if err := fPay.Await(nil); err != nil {
        return compensateAll(ctx, args.OrderID, completed, err)
    }
    completed = append(completed, StepPayment)

    fShip, err := ctx.Run(createShipment, args.OrderID)
    if err != nil {
        return SagaResult{}, err
    }
    if err := fShip.Await(nil); err != nil {
        return compensateAll(ctx, args.OrderID, completed, err)
    }
    completed = append(completed, StepShipment)

    return SagaResult{Status: "success", OrderID: args.OrderID}, nil
}

// compensateAll iterates completed steps in reverse and runs each compensation
// as a durable ctx.Run. Returns a SagaResult with status "failed" on success,
// or propagates the compensation error if a compensation itself fails fatally.
func compensateAll(
    ctx *resonate.Context,
    orderID string,
    completed []Step,
    forwardErr error,
) (SagaResult, error) {
    _ = forwardErr // log or inspect as needed

    var names []string
    for i := len(completed) - 1; i >= 0; i-- {
        step := completed[i]
        f, err := ctx.Run(compensate, CompensateArgs{Step: step, OrderID: orderID},
            resonate.RunOpts{
                Timeout: 10 * time.Minute,
                RetryPolicy: resonate.ExponentialRetry{
                    MaxAttempts: 5,
                    Base:        200 * time.Millisecond,
                    Max:         30 * time.Second,
                    Jitter:      true,
                },
            },
        )
        if err != nil {
            return SagaResult{}, fmt.Errorf("dispatch compensation for %s: %w", step, err)
        }
        if err := f.Await(nil); err != nil {
            // A compensation that exhausts retries is a real inconsistency — log and alert.
            return SagaResult{}, fmt.Errorf("compensation for %s failed: %w", step, err)
        }
        names = append(names, string(step))
    }

    return SagaResult{
        Status:      "failed",
        OrderID:     orderID,
        Compensated: names,
    }, nil
}

// compensate dispatches to the correct undo action via a switch on Step.
// Each branch must be idempotent — compensations may be retried.
func compensate(_ *resonate.Context, args CompensateArgs) (struct{}, error) {
    switch args.Step {
    case StepShipment:
        // cancelShipment(args.OrderID)
        return struct{}{}, nil
    case StepPayment:
        // refundPayment(args.OrderID)
        return struct{}{}, nil
    case StepInventory:
        // releaseInventory(args.OrderID)
        return struct{}{}, nil
    default:
        return struct{}{}, resonate.NewNonRetryable(
            fmt.Errorf("unknown saga step %q — programming error", args.Step),
        )
    }
}

// Forward leaf stubs — replace with real service calls.
func reserveInventory(_ *resonate.Context, orderID string) (struct{}, error) {
    _ = orderID
    return struct{}{}, nil
}

func chargePayment(_ *resonate.Context, orderID string) (struct{}, error) {
    _ = orderID
    return struct{}{}, nil
}

func createShipment(_ *resonate.Context, orderID string) (struct{}, error) {
    _ = orderID
    // Simulate a transient failure to trigger compensation.
    return struct{}{}, errors.New("shipment service unavailable")
}

func main() {
    r, err := resonate.New(resonate.Config{URL: "http://localhost:8001"})
    if err != nil {
        log.Fatalf("resonate.New: %v", err)
    }
    defer func() { _ = r.Stop() }()

    // Only the workflow needs registration; leaves reached via ctx.Run do not.
    placeOrderFn, err := resonate.Register(r, "placeOrder", placeOrder)
    if err != nil {
        log.Fatalf("Register: %v", err)
    }

    h, err := placeOrderFn.Run(context.Background(), "order-1", OrderArgs{OrderID: "order-1"})
    if err != nil {
        log.Fatalf("Run: %v", err)
    }
    res, err := h.Result(context.Background())
    if err != nil {
        log.Fatalf("Result: %v", err)
    }
    fmt.Printf("saga: %+v\n", res)
}

Read the full file on GitHub · 288 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2mo ago First seen · 288 lines · 86 tokens per session scan A a226f50b47b6

Subscribe to this mod's changes

resonate-saga-pattern-go is a skill published in the GitHub repository resonatehq/resonate-skills (7 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 86 tokens to every session and 2,832 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

golem-add-config-go

Adding typed configuration to a Go Golem agent with DefineConfiguredAgent and ctx.Config. Use when the user asks about agent config, configuration values, config structs, or passing settings to a Go agent.

golemcloud/golem · 46 tokens

golem-add-http-endpoint-go

Exposing a Go agent's methods over HTTP. Use when the user wants to add an HTTP endpoint/route to an agent, mount an agent under a URL path, make agent methods callable over HTTP, or map path/query/header parameters in a Go Golem project.

golemcloud/golem · 62 tokens

golem-http-router-go

Handling raw HTTP requests in Go with an HTTP router — mounting a net/http handler (ServeMux or any router), streaming request and response bodies, serving static files, and providing an OpenAPI document. Use when the user wants custom HTTP handling, to port an existing net/http service, to serve a website or files…

golemcloud/golem · 82 tokens

golem-make-http-request-go

Making outgoing HTTP requests from a Go agent. Use when the user wants an agent to call an external API/service, fetch a URL, or make an outbound HTTP/HTTPS request in a Go Golem project.

golemcloud/golem · 50 tokens

golem-mark-read-only-go

Marking Go agent methods as read-only for a side-effect-free guarantee and result caching. Use when the user wants a cacheable query method, a method that must not do remote side effects, or HTTP GET endpoints backed by a read-only method in a Go Golem project.

golemcloud/golem · 63 tokens

golem-add-cors-go

Configuring CORS on a Go agent's HTTP endpoints. Use when the user asks to enable CORS, allow cross-origin requests, or configure allowed origins for HTTP endpoints in a Go Golem project.

golemcloud/golem · 48 tokens