Tool: AWS CodePipeline

Query the AWS CodePipeline control plane to list pipelines, retrieve their full structure, and inspect their current execution state. Use this integration when a workflow needs to audit CI/CD status, gate deployments on pipeline health, or surface pipeline metadata to an Agent for analysis.

Overview

Property	Value
Type	`codepipeline`
Category	Tool — Cloud Infrastructure
Auth	AWS SigV4 (Access Key ID + Secret Access Key)

Operations

Operation	Tool ID	Description
List Pipelines	`codepipeline_list_pipelines`	List all CodePipeline pipelines in the account for the given region
Get Pipeline	`codepipeline_get_pipeline`	Retrieve the full structure (stages, actions, artifacts) of a named pipeline
Get Pipeline State	`codepipeline_get_pipeline_state`	Get the current execution state of every stage in a named pipeline

Configuration

Setting	Type	Required	Description
`awsRegion`	`string`	Yes	AWS region where your pipelines live (e.g. `us-east-1`). Determines the API endpoint (`codepipeline.<region>.amazonaws.com`).
`awsAccessKeyId`	`string`	Yes	AWS IAM access key ID used to sign requests. Treat as a secret — store in an environment variable such as `{{AWS_ACCESS_KEY_ID}}`.
`awsSecretAccessKey`	`string`	Yes	AWS IAM secret access key paired with the access key ID. Treat as a secret — store in an environment variable such as `{{AWS_SECRET_ACCESS_KEY}}`.
`name`	`string`	Conditional	The name of the pipeline. Required for Get Pipeline and Get Pipeline State; not used by List Pipelines. Can be supplied dynamically from an upstream block, e.g. `{{listPipelines.data}}`, or entered directly.

Outputs

All three operations return the same output shape:

Field	Type	Description
`data`	`json`	Raw AWS CodePipeline API response object. Content varies by operation — see the shape breakdown below.

`data` shape by operation

List Pipelines — data.pipelines is an array of pipeline summary objects, each with name, version, created, and updated fields.
Get Pipeline — data.pipeline is a pipeline structure object with name, roleArn, artifactStore, and a stages array. Each stage contains name and an actions array.
Get Pipeline State — data.stageStates is an array of stage state objects. Each entry includes stageName, inboundTransitionState, actionStates (with actionName, currentRevision, latestExecution, and entityUrl), and latestExecution with status (InProgress, Succeeded, Failed, etc.).

Example

[Starter] → [AWS CodePipeline: List Pipelines] → [AWS CodePipeline: Get Pipeline State] → [Agent: summarise deployment health]

A scheduled workflow first calls List Pipelines using {{AWS_ACCESS_KEY_ID}} and {{AWS_SECRET_ACCESS_KEY}} to enumerate all pipelines in us-east-1. A second block calls Get Pipeline State, passing the name of the target pipeline — referenced as {{listPipelines.data}} after extracting it in a Function block — to retrieve per-stage execution status. An Agent block then summarises any failed stages and posts an alert to Slack.

Tips

Least-privilege IAM — the credentials only need codepipeline:ListPipelines, codepipeline:GetPipeline, and codepipeline:GetPipelineState. Avoid using root or admin keys.
Dynamic pipeline names — use List Pipelines first to discover pipeline names programmatically, then wire {{listPipelines.data}} through a Function block to extract the desired name before passing it into Get Pipeline or Get Pipeline State.
Polling for status — to wait until a pipeline run completes, combine Get Pipeline State with a Loop block that checks stageStates[*].latestExecution.status and exits when all stages show Succeeded (or any shows Failed).
Region scoping — CodePipeline resources are region-specific. Ensure awsRegion matches the region where your pipelines were created; a wrong region returns an empty list rather than an error.

AWS CodePipeline