# Fix list: Keeper Secrets Manager From Anchor Terminal's listing at https://www.anchorterminal.com/tools/keeper-secrets-manager, the October 2026 research run, assessed 8 October 2026. Grade B, 69.4 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 Keeper Secrets Manager: 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, 20 out of 100, up to 10 more on the total Why it scored 20: No x402, MPP or L402 (0). The add-ons page shows Custom Pricing and Request a Quote for Secrets Manager, and the product page says it is licensed per user per year without a figure. Base plan prices are drawn by script and were blank to our reader (0). A 14-day business trial with no credit card, which the quick start says covers Secrets Manager (20). A person signs up, enables a role policy and creates the application and device in the vault or Commander (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, 63 out of 100, up to 6 more on the total Why it scored 63: The MCP server has 19 tools, lists metadata only and masks sensitive fields unless asked (17 of 25). Secrets are fetched by UID, by title or by folder, and Keeper notation reads a single field, with no paging (13 of 20). Throttling is an HTTP 403 with a JSON body and the Python SDK raises a typed `KeeperThrottleError`, but no documented list of error codes was found (10 of 20). No idempotency keys, and the MCP tool definitions carry no `readOnlyHint` or `destructiveHint` annotations, though writes go through a confirmation step (8 of 20). SDKs in seven languages, and a client needs only its configuration to start (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, 76 out of 100, up to 4.8 more on the total Why it scored 76: Graded with the hosted lines, because the vault is Keeper's cloud and the SDKs, CLI and MCP server are clients of it. Statuspage at statuspage.keeper.io with a Keeper Secrets Manager component and regional infrastructure components (20). Since 10 July 2026 it lists one minor incident, SMS two-factor messages on 9 September for about three hours, and one notice on EU KeeperPAM connections on 30 July, with nothing posted against Secrets Manager (28 of 30). No request limits with numbers were found in the reviewed documentation (0). The Python SDK retries a throttled call (HTTP 403 with `{"error":"throttled"}`) five times with backoff from 11 seconds and honours `retry_after`, and every SDK has an optional encrypted local cache for when the endpoint is unreachable, but there are no idempotency keys (10 of 15). The SaaS terms carry Service Level Objectives of 99.9 per cent monthly availability, with termination as the only remedy (8 of 10). Generally available (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. Schema & documentation, 71 out of 100, up to 4.7 more on the total Why it scored 71: No OpenAPI or other published contract for the Secrets Manager endpoint at `/api/rest/sm/v1`. The official MCP server defines JSON Schema inputs on all 19 tools (15 of 25). docs.keeper.io has an `llms.txt` index and serves every page as Markdown (10). The docs explain applications, devices, tokens and Keeper notation well, while MCP tool descriptions are short, such as Get a secret by UID, and don't say when to choose one tool over another (11 of 20). SDKs have typed record and field classes, and MCP inputs mark required fields, with free-form objects for record fields (11 of 15). Each SDK page has worked examples and there's a troubleshooting page, but no error reference was found (9 of 15). Each SDK has its own version tags and dated release notes on docs.keeper.io (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. Security & auth, 86 out of 100, up to 2.5 more on the total Why it scored 86: A one-time token bootstraps each device, which then signs every request with its own ECC secp256r1 key. Devices are locked to an IP address by default, can be given an access expiry, and are revoked one at a time with `secrets-manager client revoke`. An application reaches only the shared folders and records assigned to it (27 of 30). Shares are read-only unless made with `--editable`, and the MCP server asks for confirmation before create, update, delete, upload and unmasking, with an `--auto-approve` flag that turns this off (17 of 20). Secrets aren't untrusted content, and the MCP guides warn that data passed to third-party AI tools falls under those tools' practices (10). Device connect, access, record update and blocked-IP events go to the Advanced Reporting and Alerts module, which is sold as a separate add-on, and the MCP server keeps local audit logs (12 of 15). A valid `security.txt` expiring 31 December 2026, a Bugcrowd programme, and SOC 2 Type 2, ISO 27001 and FedRAMP High per the docs (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. ## 6. Transparency & trust, 78 out of 100, up to 1.9 more on the total Made of editorial 58, provenance 98. Why it scored 78: The SDKs, CLI, MCP server and agent kit are MIT on GitHub. The service is closed, under published SaaS terms that replaced the previous terms on 9 March 2026 (20 of 30). A privacy policy and a DPA are published and agree with the zero-knowledge model, where Keeper holds ciphertext. Retention is stated as for as long as an account is active, without periods (20 of 30). No deprecation policy for the SDKs or API was found. The legal pages keep an archive of earlier terms and a change summary (4 of 20). Hosting is on AWS in the US, US GovCloud, EU, Australia, Canada and Japan with the tenant fixed to its region. The DPA points to a sub-processor list in the trust centre, which is drawn by script and was unread (14 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 (9.1 of 10) - Privacy policy: read, states 7 of the 8 things a reader expects (9.3 of 10) ## 7. Maintenance & community, 92 out of 100, up to 0.7 more on the total Why it scored 92: Newest tag is the JavaScript GCP KMS storage 1.1.0 on 6 October 2026, two days before this check (30). Five tags since 15 September, among them JavaScript SDK 17.6.0, Java SDK 17.4.0 and CLI 1.5.0 (20). Three open issues, each with a reply, and 21 open pull requests (20 of 25). Seven official SDKs are current, the newest Python core being 17.3.0 from 15 June 2026. The Node MCP repository linked from the docs returned 404, and we didn't check the MCP registry (13 of 15). The 30 most recent workflow runs on master all succeeded, with dependency update runs among them (9 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. - unchecked: base plan prices on the business pricing page, which are drawn by script and were blank to our reader. Secrets Manager's own price is by quote. - unchecked: the sub-processor list, which the DPA places in the trust centre at trust.keeper.io. That page is drawn by script. - unchecked: whether the MCP server is in the official MCP registry, and the Docker Hub page for `keeper/keeper-mcp-server`. - unchecked: the troubleshooting page, so an error reference may exist that we didn't read. - No request limits with numbers were found. The only evidence of throttling is the SDK's retry code and release note. - The Node MCP repository linked from the docs returned 404, so the Node server could not be read. The Go and Docker server was graded instead. - Whether a trial account can keep Secrets Manager after 14 days without a quote. - The category's `secrets.rotate` capability is listed because Keeper sells rotation with Secrets Manager, but it runs through KeeperPAM with a gateway and its own licence. ## Weaknesses - Secrets Manager is priced by quote. The add-ons page shows Custom Pricing and Request a Quote - No request limits were found in the reviewed documentation. Throttling arrives as HTTP 403 with `{"error":"throttled"}` - No OpenAPI or documented raw HTTP use for `/api/rest/sm/v1`. The SDKs are the only supported route - The Node MCP guide links a GitHub repository that returned 404 on 8 October 2026 - Rotation needs a KeeperPAM licence and a Keeper Gateway, and access events are read in the separately sold reporting module ## 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. - Ask the vault owner for a new device on an application that holds only the folders the task needs. Use the Base64 configuration, not the first one-time token - Run commands under `ksm exec` so secrets arrive as environment variables and stay out of the model's context - Expect HTTP 403 with `{"error":"throttled"}` under load. The Python SDK retries five times from 11 seconds, so allow for long waits - A device is locked to the IP address it first connects from unless it was created with `--unlock-ip`. Check this before running from a dynamic address - Leave `--auto-approve` off on the MCP server. It removes confirmation for deletes and for unmasking values ## 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.