# Fix list: DigitalOcean MCP Server

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/digitalocean-mcp, the October 2026 research run, assessed 9 October 2026. Grade C, 60.6 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 DigitalOcean MCP Server: 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, 33 out of 100, up to 8.4 more on the total

Why it scored 33: No x402 or other machine payment. Stablecoin payment of an account's bill through Stripe is not one (0). The servers are free and product prices are public by the month, hour or unit (20). The documentation server needs no account, and App Platform, Functions and the container registry start at $0. The pricing FAQ describes a card on the account, and whether sign-up needs one was not established (10 of 20). A person signs up and approves OAuth or creates a token. Only the documentation server is keyless (3 of 20).

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. Reliability, 62 out of 100, up to 7.6 more on the total

Why it scored 62: Scored as a hosted server. status.digitalocean.com is a Statuspage with components by product and region (20). Its feed holds 25 entries from 24 July to 5 October 2026, eight of them maintenance. The incidents include the public API and control panel from a first update at 17:14 UTC on 24 August until mitigation at 00:54 on 25 August, write operations failing across several services for about two and a half hours on 6 August, account registration and Droplet creation on 8 and 9 August, and an interruption in the MKC1 region on 19 August. No component for the MCP servers was identified (5 of 30). The API's limits are published, 5,000 requests an hour and 250 a minute per token (15). The API docs describe rate-limit headers and `retry-after` on 429, and say PUT and DELETE are idempotent. There is no idempotency key for creates (10 of 15). The terms refer to Service Level Agreements, which were not read (3 of 10). The docs carry no beta label and the server is at 1.2.0, in an organisation named digitalocean-labs (9 of 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.

## 3. Security & auth, 67 out of 100, up to 5.8 more on the total

Why it scored 67: OAuth with short-lived tokens, or a personal access token with an expiry and custom scopes from about 170, revocable in the control panel. The OAuth challenge names only `read` and `write`, and the metadata allows the bearer header only (28 of 30). A Read Only token or custom scopes limit what any tool can do. The server has no read-only mode, requested in an issue open since 2 March 2026, and only the custom model delete tool asks for a confirmation argument (13 of 20). App logs, function results and documentation pages are untrusted content. The README links a third-party briefing on MCP risks and gives no mitigation guidance (3 of 15). The tokens page shows when a token was last used. No audit log of MCP calls was found (4 of 15). Valid security.txt, a bug bounty on Intigriti, SOC 2 Type II and SOC 3, and CSA STAR Level 1. The repository has no SECURITY.md file (19 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.

## 4. Schema & documentation, 66 out of 100, up to 5.5 more on the total

Why it scored 66: Tools declare typed inputs through mcp-go, and about 300 also declare an output schema checked by a test. Two issues from February 2026 about array parameters without `items` are still open, and `db-cluster-list` reads `per_page` as a Go integer in the source, which a JSON number may not satisfy (22 of 25). llms.txt and a Markdown twin of every docs page (10). The median description is 48 characters and 156 of 365 are 40 or fewer, such as "Delete a droplet". A few say how to choose between arguments (9 of 20). Required fields and defaults are set, enums are rare, and argument names mix `ID` and `Page` with `id` and `per_page` (8 of 15). Each service has a README with arguments and example queries, generated into the docs. No list of error responses was found (9 of 15). Releases are semver tags. There is no changelog file, and the 90-day release notes have no MCP entry (8 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.

## 5. Agent ergonomics, 67 out of 100, up to 5.4 more on the total

Why it scored 67: About 377 tools in all, but each hosted endpoint and the local `--services` flag load one service. Networking has 54, evaluations 47, Droplets 42 and databases 38, while 15 services have 15 or fewer (14 of 25). 62 tools take page or limit arguments, and few take filters (13 of 20). Handlers name the missing argument and pass the API's message through, and the server separates errors a model can correct from those it cannot. No documented error codes (11 of 20). Every tool carries the four MCP hints and a risk label, enforced by a test, and 308 are marked idempotent. Creates cannot be retried safely with a key (17 of 20). Few required parameters, with godo, pydo and doctl as official clients of the same API (12 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.

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

Why it scored 70: v1.2.0 was tagged on 7 October 2026 and the npm package is at the same version (30). Eight tags since 23 July 2026 (20). 193 commits in 90 days. Eight open issues and 23 open pull requests. The credential issue of 26 May 2026 has two comments, neither from a maintainer, and the read-only request of 2 March 2026 is open (12 of 25). The official MCP registry returned one third-party server for the search digitalocean and none from DigitalOcean (0). Dependabot runs weekly, and CI lints, checks formatting and runs tests with the race detector on Linux and macOS. The workflow's result was not seen (8 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.

## 7. Transparency & trust, 87 out of 100, up to 1.1 more on the total

Made of editorial 74, provenance 100.

Why it scored 87: The server is MIT and the platform is closed with clear terms (28 of 30). A privacy policy and a DPA are published. The DPA promises deletion of personal data within 30 days of account deactivation. Nothing found says what the hosted MCP servers log, and the source can ship tool-call logs to a WebSocket endpoint when configured (20 of 30). The `alert-policy-*` tools carry a deprecation notice with no date, the terms promise as much notice of changes as is reasonably practicable, and the April 2026 service rename had no notice we found (9 of 20). Sub-processors are listed with locations on a page dated 10 June 2026, with 30 days to object to a new one (17 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.

## Deductions

Each comes off the total. A fixed and documented problem counts for less at the next check.

- 26 May 2026, still open on 9 October 2026. An issue reports that `db-cluster-list`, `db-cluster-get` and the database user tools return passwords in plaintext, so a lookup puts credentials in the model's context. The handlers at v1.2.0 return the API's object with no redaction. A token without `database:view_credentials` limits it. 3 points (https://github.com/digitalocean-labs/mcp-digitalocean/issues/372).

## 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 hosted servers themselves. The request for robots.txt on docs.mcp.digitalocean.com answered 302 to a docs page, which the brief's rule does not cover, so no `initialize` or `tools/list` was sent. Tool counts are from the source at v1.2.0 and may differ from what the hosted endpoints serve.
- unchecked: the Service Level Agreements the terms refer to. No SLA page was read.
- unchecked: whether sign-up needs a payment method, and the current free credit for new accounts.
- unchecked: the result of the CI workflow on the default branch, and the GitHub releases page. Release dates are from git tags.
- unchecked: whether DigitalOcean keeps an audit log of API or MCP calls. The team security history docs were not read.
- unchecked: the OAuth authorisation server's metadata at cloud.digitalocean.com, including PKCE, dynamic client registration and token lifetime.
- unchecked: security incidents at DigitalOcean in the last 12 months beyond the status feed. No advisory list was searched.
- Whether the old `gradient-ai` hosted address still answers after the 23 April 2026 rename, and whether users were told, was not established. No deduction was taken.
- The password exposure in the database tools is the issue reporter's account plus our reading of the handlers. We did not call the tools, and did not read the godo type definitions.
- The Acceptable Use Policy forbids harvesting or scraping content of the Services. This matters before any probe is run.
- The lead was right on the vendor, the repository, the licence and the hosted and local servers. The docs list a 23rd endpoint, Knowledge Bases, on another domain.

## Weaknesses

- No read-only mode on the server. A request for one has been open since 2 March 2026.
- `db-cluster-list` and `db-cluster-get` return the API's database object unredacted. An issue reporting plaintext passwords has been open since 26 May 2026.
- The status feed lists several long incidents since 24 July 2026, including the public API and control panel on 24 and 25 August.
- Many descriptions are one short phrase, such as "Delete a droplet", and argument names mix `ID`, `Page` and `per_page` styles.
- No DigitalOcean entry was found in the official MCP registry, and no changelog file or MCP release note was found.

## 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.

- Add only the service endpoints the task needs. Networking has 54 tools, evaluations 47 and Droplets 42.
- Pass `--services` to the local server. With no value it loads every service, about 377 tools.
- Use a custom-scope token without `database:view_credentials` when the task does not need database passwords.
- Use the `insights-alert-rule-*` tools. The `alert-policy-*` tools are marked deprecated and will be removed.
- Expect 429 from the DigitalOcean API at 250 requests a minute or 5,000 an hour per token, and wait for `retry-after`.

## 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.