Optimizely Experimentation

# Fix list: Optimizely Experimentation

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/optimizely, the October 2026 research run, assessed 8 October 2026. Grade B, 62.6 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 Optimizely Experimentation: 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, 0 out of 100, up to 12.5 more on the total

Why it scored 0: No x402, MPP or L402 (0). No price is published. The plans page says every plan is individually packaged and leads to a demo request (0). No free plan or trial was found on the plans page, the product pages or the docs (0). Access needs a subscription and a browser sign-in through Opti ID (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. Security & auth, 66 out of 100, up to 6 more on the total

Why it scored 66: The MCP server uses OAuth 2.0 with PKCE (S256), dynamic client registration and a revocation endpoint, with one scope, `mcp:tools`. REST personal access tokens are revocable, named and sent in a header, and administrators can revoke any collaborator's token. The REST OAuth flow has the single scope `all`, described as full access to the account (22 of 30). Sessions and tokens carry the user's role, Feature Experimentation has granular roles for flags, environments and audiences, and change approvals have routed flag changes to reviewers since 6 August 2026. No read-only mode for the MCP server was found, and confirmation before a write is left to the client (13 of 20). No guidance on prompt injection was found, though tools return names, descriptions and hypotheses that people wrote (3 of 15). Change history for projects and flags is in the application, at `/v2/changes` and in webhooks (12 of 15). ISO 27001, ISO 27017, ISO 27018, SOC 2 Type 2 and PCI DSS are stated for Web and Feature Experimentation, with annual penetration tests and a bug bounty programme reachable at securityeng@optimizely.com. No security.txt and no public programme link (16 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.

## 3. Reliability, 73 out of 100, up to 5.4 more on the total

Why it scored 73: Graded on the hosted service with the hosted lines. Statuspage at status.optimizely.com with 174 components, among them Optimizely REST API, Results API and Events API for Feature Experimentation and Web Experimentation (20). No incident was filed against an Experimentation component between 10 July and 8 October 2026. On 23 July 2026 sign-in through Optimizely Identity failed with 429 errors for about 3 hours 15 minutes because of a disruption at Okta, marked major, and on 17 September 2026 a 3-minute minor incident sent users to the wrong SSO page. Both affect the OAuth sign-in the MCP server depends on (20 of 30). The Feature Experimentation API is limited to 2 requests a second and 120 a minute, the Optimizely API to 100 a minute and the results endpoints to 20 a minute. No limit was found for the MCP server (13 of 15). No 429 response, `Retry-After` header, backoff guidance or idempotency key was found in the v2 spec or the pages read (0 of 15). The Service Level Agreement V 2024-05 gives credits from 10 per cent once downtime passes 43 minutes in a month, capped at 50 per cent (10). The REST APIs are versioned and the MCP pages carry no beta 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. Agent ergonomics, 68 out of 100, up to 5.2 more on the total

Why it scored 68: The MCP server lists seven tools, which is 25 on the checklist. We took 2 off because the tool definitions were unread and their size is unknown (23 of 25). `per_page` up to 100 and `page` on the v2 lists, `page_token`, `sort` and `query` on the flags list, and time windows on results (17 of 20). The Optimizely API returns `code`, `message` and a `uuid`, and the Feature Experimentation API returns a problem detail object. MCP error results were unread (13 of 20). No idempotency keys were found and the tool annotations were unread. The docs say the agent confirms project, environment and details before a change, which is client behaviour, and entities are archived through the lifecycle tool (5 of 20). The MCP server needs only a URL. The official SDKs in eleven languages evaluate flags and log events and do not cover the management APIs (10 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.

## 5. Schema & documentation, 74 out of 100, up to 4.2 more on the total

Why it scored 74: Public Swagger 2.0 spec for the Optimizely API with 65 paths and 93 operations. The spec for the Feature Experimentation API at `/flags/v1` answers 401 without a token, though its reference pages carry the same structured definitions. We could not read the MCP tool schemas, which need a sign-in (20 of 25). llms.txt with 11,441 lines and a Markdown copy of every docs page (10). All 93 operations have a description. The MCP overview gives each of the seven tools one line of purpose and no guidance on when not to use it (12 of 20). 111 enums and 193 required lists in the v2 spec. The MCP server puts queries and writes behind two general tools, `exp_execute_query` and `exp_manage_entity_lifecycle`, with schemas and templates fetched by two more (10 of 15). 274 examples in the v2 spec, with 401 on 77 operations, 403 on 73, 404 on 59 and 400 on 41, and no 429 (11 of 15). Versions are in the path (`/v2`, `/flags/v1`) and each product has dated release notes with an RSS feed. No changelog for the APIs alone was found (11 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.

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

Made of editorial 63, provenance 86.

Why it scored 75: Closed hosted service under a published Online Software Subscription Agreement and API, SDK and MCP terms. The SDKs are open source under Apache 2.0 (17 of 30). The privacy notice of 1 July 2026 says it does not cover customer data and points to the subscription agreement and the Data Processing Agreement (V 2026-01), which agree. Event data is kept 18 months, and customer data may be deleted 30 days after a subscription ends. How the MCP server and Opal handle prompts for Experimentation is not stated on the MCP pages (22 of 30). The API, SDK and MCP terms let Optimizely revise, suspend or cancel any API or MCP at any time under its documentation and policies. An SDK compatibility matrix exists. No deprecation policy with a notice period was found for the APIs (6 of 20). The sub-processor list, updated 14 September 2026, names providers by product with hosting locations. Web and Feature Experimentation list Amazon Web Services in the United States and Google Cloud in the United States and Europe (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 5 of the 7 things a reader expects, and has 1 clause that costs points (6.3 of 10)
- Privacy policy: read, states 7 of the 8 things a reader expects (9.3 of 10)
- security.txt: not found (0 of 10)

## 7. Maintenance & community, 78 out of 100, up to 1.9 more on the total

Why it scored 78: `@optimizely/optimizely-sdk` 6.6.1 was tagged on 18 September 2026, and the newest product release notes are dated 3 September 2026 for Feature Experimentation and September 2026 for Web Experimentation (30). Feature Experimentation notes on 22 July, 6 August and 3 September 2026, plus SDK releases 6.6.0 and 6.6.1 (20). Closed service with public release notes, an RSS feed and a support centre. We did not read how quickly questions are answered (9 of 15). Official SDKs are current in eleven languages. The MCP server is not in the official MCP registry, where a search for optimizely returns nothing (12 of 15). The JavaScript SDK repository has six CI workflow files and 8 commits since 10 July 2026 (7 of 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.

## 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.

- The lead was right that the docs hub lists REST APIs and an MCP server. The hub is at docs.optimizely.com, and docs.developers.optimizely.com redirects there
- The API, SDK and MCP terms (clause 1.11) forbid a licensee from disseminating performance information about any API, SDK or MCP, including uptime, response time and benchmarks. This matters for any probe results we publish and wants a decision before probes run
- provenance.terms points at the Online Software Subscription Agreement, which is an HTML page. The API, SDK and MCP terms are part of it and are published only as a PDF
- provenance.privacy points at the privacy notice, which says it does not apply to customer data. Customer data is governed by the Data Processing Agreement, published as a PDF
- unchecked: the live `tools/list` response of the MCP server, its input schemas and annotations, which need a sign-in. Tool names and purposes come from the docs overview
- unchecked: the spec for the Feature Experimentation API at `/flags/v1`, which answers 401 without a token. Its operations were read from the reference pages
- unchecked: whether the REST APIs return 429 with `Retry-After`. The v2 spec documents no 429 and we did not search every docs page
- unchecked: GitHub stars and issue response times. We read the JavaScript SDK repository by clone only
- unchecked: the bug bounty programme's platform and scope. The security page names a programme and an email address without a link
- No free plan or trial was found on the plans page, the experimentation and feature management product pages or the docs index. An older free feature flagging page returns 404
- The 23 July 2026 identity incident was caused by Okta and was not filed against an Experimentation component. We counted it as minor for this listing because existing tokens were not shown to be affected
- Capabilities leave out events and funnels. The Events API and the SDKs log events, but the management surface an agent uses has no funnel or retention query

## Weaknesses

- No price is published. The plans page says every plan is individually packaged and leads to a demo request
- No free plan or trial was found, and the MCP server needs an account with Opal enabled
- No 429 response, `Retry-After` header or idempotency key was found in the spec or the pages read
- The REST OAuth flow has one scope, `all`, described as full access to the account
- The spec for the Feature Experimentation API at `/flags/v1` answers 401 without a token, and the MCP server is not in the official MCP registry

## 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.

- Connect to https://exp.mcp.opal.optimizely.com/mcp and complete OAuth in a browser. The session is time-limited, so expect to sign in again
- Call `exp_get_schemas` before `exp_execute_query`, and `exp_get_entity_templates` before `exp_manage_entity_lifecycle`, to learn the fields each entity needs
- Confirm the project and environment with the user before any create or update. Changes go to live Optimizely data
- Use the Visual Editor for Web Experimentation variation code. The MCP server does not write custom HTML, CSS or JavaScript
- Over REST, stay under 2 requests a second on `/flags/v1`, 100 a minute on `/v2` and 20 a minute on the results endpoints

## 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.