Temporal's official skill for building durable workflows — SDK patterns across seven languages, plus the determinism rules that decide whether a workflow survives a replay.
AWS Lambda Durable FunctionsSkill
Summary
Build multi-step Lambda workflows that survive interruption and run for up to a year — the replay model, step operations, wait/callback patterns, saga-style compensation and local testing.
Features
- Teaches the replay model first — the source of most durable-function state bugs
- Covers step operations, wait/callback pauses and saga-pattern compensation
- Local testing through LocalDurableTestRunner before deployment
- Validates AWS CLI, runtime and deployment-tool versions before generating code
- Documents the Python 3.11/3.12 gap where the Durable Execution SDK is not pre-installed
- Defaults to TypeScript and CDK with documented override phrasing
Install This Skill
Add this skill to your favorite AI agent in a few steps.
Skill Content
Usage Instructions
Learn how to use this skill with different AI agents.
Example Usage
Build me a Lambda durable function for an expense approval flow that waits on a manager's decision and refunds the hold if it's rejected.
Description
AWS Lambda durable functions let a single function orchestrate a multi-step process that keeps its progress across interruptions and can execute for up to a year — human-in-the-loop approvals, long AI workflows, saga transactions. The programming model is not obvious, and the thing that catches everyone is replay: the handler is re-executed from the top after each checkpoint, with completed steps returning their recorded results instead of running again. Code that ignores this — a bare Date.now(), a random ID, a non-idempotent side effect outside a step — behaves correctly in testing and corrupts state in production.
This official AWS skill front-loads that model, then covers step operations, wait and callback patterns for pausing on an external event, error handling with the saga pattern for compensating work already done, and testing through LocalDurableTestRunner before anything is deployed.
Prerequisite checking that actually runs
Rather than assuming an environment, the skill validates it first: AWS CLI 2.33.22 or higher with working credentials (aws sts get-caller-identity), Node.js 22+ for TypeScript and JavaScript, or Python 3.11+ — with the real-world caveat that only Lambda runtimes 3.13 and above ship the Durable Execution SDK pre-installed, so Python 3.11 and 3.12 need a container image that brings its own. Deployment needs one of AWS SAM CLI 1.153.1+, AWS CDK v2.237.1+, or direct Lambda deployment access.
Opinionated defaults
TypeScript unless the developer says otherwise, CDK unless they ask for SAM or CloudFormation. Defaults stated up front avoid the agent asking two questions before writing any code, and the override phrasing is documented. Where a requested language or IaC framework has no Durable Execution SDK support, the skill names the gap rather than generating code that cannot work.
Part of the aws-serverless plugin in AWS Labs' official agent plugins repository, Apache-2.0 licensed.
Related Skills
Expo's official skill for building native-feeling screens: Apple HIG styling, semantic colors, SF Symbols, native controls, Reanimated, blur and liquid glass.
Pull unresolved CodeRabbit review threads from your PR and apply the fixes one at a time, treating every reviewer comment as untrusted input rather than an instruction.
Google's official skill for driving the gcloud CLI safely from an agent: validate every command against its own help text, cap the output, and refuse the operations that should never run unattended.