AWS S3 MCP integration for AI agents.
Connect AI agents to AWS S3 through 7 structured actions, including download object, delete object, and generate presigned url. Review authentication, inputs…
- ACTION CONTRACTS
- 7
- AUTHENTICATION
- Custom credentials
- SOURCE STATUS
- Provider-backed
CATALOG SOURCE REVIEWED AUGUST 23, 2026 / ACTION NAMES AND SCHEMAS DERIVED FROM WORKING MACHINES PROVIDER SOURCE
CAPABILITY PROFILE
What the AWS S3 integration exposes
The AWS S3 Agent App exposes 7 provider-backed actions for storage and developer tools work. Its current contract lets an authorized agent download one s3 object into local transit file storage, delete one s3 object, and generate a pre-signed s3 url for reading, uploading, or deleting one object. These operations are called through Working Machines as typed capabilities rather than through browser navigation or copied UI steps.
A connection uses Custom credentials. Before execution, the agent can inspect the selected action, its required fields, declared scopes, and expected output contract. Provider credentials remain inside the Working Machines runtime; the calling agent receives the capability and its structured result, not the underlying secret.
For reliable operation, start with the narrowest action that satisfies the task, resolve stable provider identifiers before changing state, and validate the returned object or status after execution. Availability still depends on the connected AWS S3 account, granted provider permissions, workspace policy, region, plan, and upstream API behavior.
VERIFIED ACTION SAMPLE
Real AWS S3 capabilities.
Showing 7 of 7 actions. Risk labels are conservative signals based on operation names, not substitutes for provider documentation or runtime policy.
download_objectREADDownload one S3 object into local transit file storage.
INPUTS: bucket / objectKey / region / versionId / fileName
delete_objectHIGH IMPACTDelete one S3 object.
INPUTS: bucket / objectKey / region / versionId
generate_presigned_urlREVIEWGenerate a pre-signed S3 URL for reading, uploading, or deleting one object.
INPUTS: bucket / objectKey / region / method / expiresSeconds
head_objectREVIEWFetch structured metadata for one S3 object.
INPUTS: bucket / objectKey / region / versionId
list_bucketsREADList Amazon S3 buckets visible to the connected AWS credential.
INPUTS: region / prefix / marker / maxKeys
list_objectsREADList objects in an S3 bucket with the ListObjectsV2 API.
INPUTS: bucket / region / prefix / delimiter / continuationToken
put_objectREVIEWUpload one object to S3 from a public URL, plain text, or base64-encoded content.
INPUTS: bucket / objectKey / region / sourceUrl / contentText
INPUT CONTRACTS
Know what the action needs before it runs.
download_object
bucketOPTIONAL- The S3 bucket name.
objectKeyREQUIRED- The complete S3 object key. Slashes are preserved as key delimiters.
regionOPTIONAL- The AWS region, for example `us-east-1`. Connections can omit it on each call to reuse the connected default region.
versionIdOPTIONAL- The optional object version ID.
fileNameOPTIONAL- An optional filename override for the local transit file.
delete_object
bucketOPTIONAL- The S3 bucket name.
objectKeyREQUIRED- The S3 object key.
regionOPTIONAL- The AWS region, for example `us-east-1`. Connections can omit it on each call to reuse the connected default region.
versionIdOPTIONAL- The optional object version ID.
generate_presigned_url
bucketOPTIONAL- The S3 bucket name.
objectKeyREQUIRED- The S3 object key.
regionOPTIONAL- The AWS region, for example `us-east-1`. Connections can omit it on each call to reuse the connected default region.
methodOPTIONAL- The HTTP method that the signed URL should allow.
expiresSecondsOPTIONAL- How long the signed URL remains valid, in seconds.
head_object
bucketOPTIONAL- The S3 bucket name.
objectKeyREQUIRED- The S3 object key.
regionOPTIONAL- The AWS region, for example `us-east-1`. Connections can omit it on each call to reuse the connected default region.
versionIdOPTIONAL- The optional object version ID.
PROVIDER-SPECIFIC WORKFLOWS
Jobs this Agent App can support
Inspect Download Object
Download one S3 object into local transit file storage. Use this as a bounded discovery step, retain the returned identifier, and avoid expanding the read beyond the task's stated scope.
download_objectControl Delete Object
Delete one S3 object. Resolve the target first, present material changes for confirmation, and make retries idempotent where the provider supports it.
delete_objectVerify with List Buckets
List Amazon S3 buckets visible to the connected AWS credential. Compare the returned provider state with the intended outcome and preserve stable IDs or canonical links in the run record.
list_bucketsSAFETY BOUNDARY
Operate AWS S3 with explicit limits
- Authorize AWS S3 with Custom credentials and grant only the provider access required by the selected actions.
- 1 action is change-capable by name. Confirm the target identity and material parameters before allowing a write.
- Treat `delete_object` as high-impact operations and require an explicit approval boundary.
- No provider scope string is declared on the sampled actions. Verify the connected account's actual permissions in AWS S3 rather than assuming unrestricted access.
- After a call, inspect the structured result and execution record before reporting that the AWS S3 task completed successfully.
CONNECTION MODEL
Custom credentials
AWS S3 uses Custom credentials. Connection requirements and availability are determined by the provider and the Working Machines deployment.
Working Machines stores provider credentials behind the execution boundary. An agent can use an authorized connection identity, but catalog discovery alone does not reveal OAuth tokens, API keys, or provider secrets.
- READ SIGNALS
- 3
- WRITE SIGNALS
- 0
- HIGH IMPACT
- 1
- REVIEW SIGNALS
- 3
EVIDENCE AND AVAILABILITY
Provider reference
Action names, input fields, authentication types, and counts on this page are generated from the Working Machines provider catalog. Provider behavior, quotas, object semantics, account eligibility, and regional availability remain governed by AWS S3.