Amazon Bedrock AgentCore Identity

# Fix list: Amazon Bedrock AgentCore Identity

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/agentcore-identity, the October 2026 research run, assessed 8 October 2026. Grade BB, 74.8 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Amazon Bedrock AgentCore Identity: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Payments & pricing, 30 out of 100, up to 8.8 more on the total

Why it scored 30: No x402, MPP or L402 for paying AWS. AgentCore payments is a separate capability for agents paying third parties (0). $0.010 per 1,000 token or API key requests on the public pricing page (20). No Identity free tier, but new accounts get up to $200 of Free Tier credit and the FAQ says most new customers need no payment method, while AWS may still ask for one, so half, as on our other AWS listings (10). A person creates the AWS account and registers an OAuth client with each provider (0).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 2. Agent ergonomics, 76 out of 100, up to 3.9 more on the total

Why it scored 76: Responses are small, a token or an authorisation URL with a session status, and list calls take `maxResults`. There is no MCP server for Identity to weigh (20 of 25). List operations page with `nextToken` and `maxResults`, and no filters were found (15 of 20). Named exceptions with recovery advice, and a missing consent comes back as `authorizationUrl` and `sessionUri` in a 200 response, which an agent can act on (18 of 20). Token reads are safe to repeat and `forceAuthentication` restarts consent, but no client token for idempotent creates was found and there are no MCP annotations (10 of 20). AWS SDKs in nine languages plus AgentCore SDKs for Python and TypeScript with `@requires_access_token` and `@requires_api_key` decorators. `GetResourceOauth2Token` has four required members, and user-delegated flows need the owner to host a callback endpoint (13 of 15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 3. Reliability, 85 out of 100, up to 3 more on the total

Why it scored 85: AWS Health Dashboard with a per-service, per-Region feed for Bedrock AgentCore (20). The dashboard's history file, read on 8 October 2026, lists one event naming AgentCore in the last 90 days, elevated packet loss in one Availability Zone of eu-south-2 on 4 October 2026 from 8:48 to 11:28 AM PDT, at informational status and shared with 31 other services. The us-east-1 feed had no items. Minor only (20 of 30). Quotas published per operation, 200 requests a second for the workload access token calls and 20 for management calls (15). The API reference documents `ThrottlingException` (429) and `InternalServerException` (500) with advice to retry with exponential backoff, but no idempotency token was found on `CreateWorkloadIdentity` (12 of 15). The AgentCore FAQ says the Amazon Bedrock SLA applies, 99.9 per cent a Region, though that SLA's definitions speak of model APIs and do not name AgentCore (8 of 10, our call). Generally available since October 2025, and the consent portal carries no preview label (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 4. Security & auth, 84 out of 100, up to 2.8 more on the total

Why it scored 84: IAM with SigV4 and short-lived role credentials, plus a workload access token that carries the agent's identity and the user's, and policies that name one workload identity and one credential provider ARN (30). Users consent through the provider's own OAuth screen with session binding, but AWS says the service enforces no binding between workload identities and credential providers beyond the owner's IAM policy, the user ID route is unverified, and no call to revoke one user's grant was found (15 of 20). The service returns tokens and keys, not untrusted content (10). The guide points to CloudTrail for `GetWorkloadAccessTokenForUserId` calls and KMS signing, but no Identity-specific CloudTrail page listing logged events was found, unlike Gateway and Agent Registry (11 of 15). A disclosure programme on HackerOne, public bulletins, AgentCore in SOC 1, 2 and 3 scope on the list of 11 August 2026 and ISO 27001 per the compliance page, with the security.txt expired since 24 September 2026 and no paid bounty found, read as our other AWS listings read it (18 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 5. Maintenance & community, 70 out of 100, up to 2.6 more on the total

Why it scored 70: The newest Identity change is the consent portal, announced 1 September 2026, 37 days before the check (20 of 30). In the last 90 days we could date the consent portal and Python SDK 1.21.0 of 6 August 2026, which added workload access token propagation. Private Key JWT is listed under July without a day. The Python SDK had ten releases in the 90 days to 8 October across all of AgentCore (15 of 20). Public release notes and What's New, with AWS Support and re:Post, and issue triage workflows in the SDK repository. We did not read the issue tracker (10 of 15). Current official SDKs, `bedrock-agentcore` 1.24.1 on PyPI on 7 October 2026 and `@aws-sdk/client-bedrock-agentcore` 3.1148.0 (15). The SDK repository runs CI, integration tests, a breaking-change check, security scanning and Dependabot (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## 6. Transparency & trust, 75 out of 100, up to 2.2 more on the total

Made of editorial 61, provenance 88.

Why it scored 75: Closed service under the AWS Service Terms updated 1 October 2026, with Apache-2.0 SDKs (15 of 30). The Identity data protection pages state KMS encryption at rest with an optional customer managed key and warn that names and free-text fields can reach diagnostic logs. The privacy notice is dated 18 May 2026. No retention period for stored tokens or request metadata was found (20 of 30). No deprecation policy for the service found. The one dated notice we saw is the starter toolkit's deprecation on 27 March 2026, in a security bulletin (8 of 20). A sub-processor list updated 28 July 2026 and regional endpoints in 21 Regions, with the vault held in the Region the owner picks (18 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

Provenance checks not met in full (half of this category, computed from checked facts):

- Terms of service: read, states 6 of the 7 things a reader expects, and has 3 clauses that cost points (3.1 of 10)
- security.txt: published but past its Expires date (5 of 10)

## 7. Schema & documentation, 88 out of 100, up to 2 more on the total

Why it scored 88: Service models for `bedrock-agentcore` and `bedrock-agentcore-control` ship in the AWS SDKs. We confirmed the SDK clients and the API reference, and did not open the model files (25). llms.txt for the developer guide and for the API reference, with a Markdown twin of every page (10). API reference descriptions are one line each, while the guide says when to use the JWT route and when the user ID route (15 of 20). Typed members with enums, patterns and length limits, such as `oauth2Flow` and `credentialProviderVendor`, and one string map, `customParameters` (14 of 15). Named errors with HTTP codes on every operation and CLI and Python examples in the guide, but no examples on the API reference pages we read (12 of 15). Release notes by month without days, SDK changelogs with dates, and the notes' RSS feed was not found at the address we tried (12 of 15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- unchecked: the amazon.com registration date (1994-11-01) is carried from our other AWS listings, because WHOIS was not reachable from this session.
- unchecked: the SDK service model files and the GitHub issue trackers were not read, and GitHub stars were not fetched.
- unchecked: the AWS Customer Agreement, the DPA and the sub-processor table's rows were not re-read today. Only the sub-processor page's date was confirmed.
- The day in July 2026 on which Private Key JWT launched was not established. The release notes are dated by month.
- No per-user grant revocation call and no Identity-specific CloudTrail page were found in the pages read. Either may exist elsewhere in the IAM or CloudTrail documentation.
- The Bedrock SLA's definitions refer to model APIs. AWS's FAQ says it applies to AgentCore, and how a claim for Identity would be measured is not stated.
- The lead was accurate. One point to add is that the consent portal needs an AgentCore Gateway and an OIDC sign-in provider.

## Weaknesses

- No operation to revoke or delete one user's stored grant was found. `forceAuthentication` clears a refresh token, and AWS says it cannot detect a revocation made at the provider.
- The consent portal, launched 1 September 2026, attaches to one AgentCore Gateway with JWT inbound auth and cannot use GitHub, Slack, Salesforce, Atlassian or LinkedIn as its sign-in provider.
- `GetWorkloadAccessTokenForUserId` takes a user ID string the platform does not verify, so the binding to a user rests on the caller and its IAM policy.
- AWS states the service enforces no binding between workload identities and credential providers in one account beyond the IAM policy the owner writes.
- No CloudTrail page for AgentCore Identity was found in the developer guide, though Gateway and Agent Registry each have one.

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Get a workload access token first (`GetWorkloadAccessTokenForJWT` in production), then pass it as `workloadIdentityToken` to `GetResourceOauth2Token` or `GetResourceApiKey`.
- When `GetResourceOauth2Token` returns `authorizationUrl` instead of `accessToken`, send the URL to the user and call again with the same `sessionUri` after consent.
- For user-delegated flows, host an HTTPS callback, register it with `UpdateWorkloadIdentity` as an allowed return URL, and call `CompleteResourceTokenAuth` after checking the user's session.
- Ask for refresh tokens in the provider's own way, such as `access_type=offline` in `customParameters` for Google or the `offline_access` scope for Microsoft and Atlassian.
- Treat a returned token as possibly revoked. On a 401 from the resource server, retry with `forceAuthentication` set to true.

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.