Dropbox Sign

# Fix list: Dropbox Sign

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/dropbox-sign, the October 2026 research run, assessed 7 October 2026. Grade B, 68.9 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 Dropbox Sign: 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, 25 out of 100, up to 9.4 more on the total

Why it scored 25: Read with the hosted rubric. No x402, MPP or L402 (0). Plan prices are public, Essentials at $900 a year from 50 requests a month and Standard at $3,000 a year from 100, with larger volumes on a selector and Premium by quote. That is plan pricing, not a per-request price (10 of 20). Test mode is free on every endpoint from a free account, but its requests aren't legally binding, and the pages we read don't say whether signup needs a card (15 of 20). A person has to create the account and copy the key from the web app, and production needs a paid plan (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, 65 out of 100, up to 6.1 more on the total

Why it scored 65: API keys go in an HTTP Basic header, up to four an account for rotation, and each has full access to the account. OAuth 2.0 tokens carry seven scopes but serve apps acting for other users and need support approval. Scored between plain revocable keys and scoped OAuth (25), less 5 because the docs show the key inside the URL as `https://KEY:@api.hellosign.com`, a judgement call since that is URL userinfo, not a query string (20 of 30). The limited OAuth scopes and test mode narrow what a call can do, but there is no read-only key and no confirmation step on cancel, remove or template delete (9 of 20). Responses can carry signer-entered field values and decline reasons, and no injection guidance was found (5 of 15). The API Dashboard records the account's API requests, responses and callbacks, live and test, and each completed document carries an audit trail with timestamps, IP addresses and a SHA-256 hash (13 of 15). SOC 2 Type II, ISO 27001 and ISO 27018 at trust.dropbox.com, a bug bounty and disclosure programme on Intigriti, and webhook HMAC verification. No standard security.txt on the Sign hosts (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.

## 3. Reliability, 75 out of 100, up to 5 more on the total

Why it scored 75: Read with the hosted lines and scored on the public REST API at api.hellosign.com, the surface an agent would call. status.hellosign.com on Statuspage lists 13 components in four groups with incident history (20). Two incidents in the 90 days to 7 October 2026, both marked minor, a Salesforce integration fault on 21 August (6 h 43 min) and file upload problems on 31 August (54 min). The last major incident was an outage on 2 January 2026 (3 h 12 min), outside the window (20 of 30). Rate limits with numbers, 100 requests a minute standard, 25 on 17 higher-tier endpoints and 10 in test mode, although the header table describes the limit as hourly (15). The docs describe 429 with `X-Ratelimit` headers, tell clients to respect Retry-After and back off, and the error catalogue marks each error retryable or not. No idempotency keys or safe-retry guidance for sends were found (10 of 15). The terms supply the service as is and as available, and no SLA was found (0). The API is at v3 and 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. Agent ergonomics, 72 out of 100, up to 4.6 more on the total

Why it scored 72: List calls take `page_size` from 1 to 100 (default 20), and files are fetched by separate calls as binary, data URI or URL, so status checks stay small. No field selection was found (15 of 25). Page-number pagination with totals in `list_info`, and a `query` search language over title, signer, sender, dates, completion state and metadata. The docs warn of a short indexing delay after creation (18 of 20). Errors carry `error_name` and `error_msg`, with 20 documented names, a remediation line each and a retryable flag, plus warnings in successful responses (20). No idempotency keys in the docs or spec, so a retried send can create a duplicate. Cancel returns 200 when queued and confirms by event. The retryable flags are the only safe-retry aid (4 of 20). A template send needs only `template_ids` and `signers`, `test_mode` is a single flag, and official SDKs cover six languages (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. Transparency & trust, 68 out of 100, up to 2.8 more on the total

Made of editorial 51, provenance 85.

Why it scored 68: The service is closed under terms effective 7 January 2025. The spec repository is Apache 2.0 and the SDKs are MIT (15). The terms let Dropbox delete customer data any time after 30 days from termination. The data processing agreement, dated 25 October 2021, promises deletion within a commercially reasonable period on request and still points to a hellosign.com sub-processor address. The privacy policy of 14 January 2025 gives no retention period. The documents agree but stay vague (18 of 30). Prior SDK major versions are patched for 12 months and the API returns a `deprecated_parameter` warning, but no API deprecation policy with notice periods was found (8 of 20). The privacy policy links a Sign sub-processor list and the agreement promises advance notice of new sub-processors with a 60-day objection window. The list itself returned only its title to our reader, so names and locations are unconfirmed. Data residency is a Premium option (10 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 2 clauses that cost points (5.1 of 10)
- security.txt: not found (0 of 10)

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

Why it scored 88: One public OpenAPI 3.0.3 spec in hellosign/hellosign-openapi with 68 paths, 74 operations and 227 schemas, which also drives the docs and SDKs (25). llms.txt at developers.hellosign.com, a Markdown copy of each page by appending .md, and scoped llms-full.txt files by section and language (10). Operation descriptions average about 220 characters and several state consequences, such as cancel being irreversible and asynchronous. A glossary written for models covers core objects, signers, fields and workflows. Few descriptions say when not to use an endpoint (14 of 20). 67 enums, 239 required lists and about 80 length or range constraints. Sends are multipart forms with indexed keys, and `metadata` and `custom_fields` are loosely typed (11 of 15). Request and response examples with SDK samples in six languages, and a catalogue of 20 HTTP error names with cause, remediation and a retryable flag, embedded in the spec as `x-error-codes` (15). The path carries v3 and a dated changelog has 48 entries since November 2022. No written API versioning policy was found beyond the SDK one (13 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.

## 7. Maintenance & community, 85 out of 100, up to 1.3 more on the total

Why it scored 85: SDK version 1.13.0 reached npm and PyPI on 10 September 2026, 27 days before the check, and the spec repository had commits on 6 October, among them a new Document Field Detection endpoint (30). Dated changelog entries on 31 July, 19 August and 3 September 2026, plus SDK 1.12.0 and 1.13.0 on 10 September (20). A closed service with a public changelog, an API support address (apisupport@hellosign.com), a help centre and public GitHub issues, nine open on the spec repository. We didn't test response times (11 of 15). Current official SDKs for Python, Node, PHP, Java, Ruby and .NET, all pushed in September 2026 (15). CI in the spec repository builds the spec and each SDK and fails on uncommitted generated code, and Node dependencies were upgraded on 6 October (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: the Sign sub-processor list at www.dropbox.com/privacy/subprocessor/sign, which returned only its title to our reader, so names and locations are unconfirmed.
- unchecked: PyPI weekly downloads for dropbox-sign. pypistats.org refused us for a rate limit.
- unchecked: whether creating a free account for test mode needs a card. The docs and pricing page don't say and we didn't sign up.
- unchecked: prices at the 250, 500 and 750+ steps of the pricing selector, and the month-to-month price. The static page showed $75 and $250 a month beside $900 and $3,000 a year.
- unchecked: documents in the trust centre beyond the public summary, such as the SOC 2 report and penetration test summary.
- The rate-limit docs give per-minute limits, describe `X-Ratelimit-Limit` as hourly and show an example message of 2,000 an hour. Which window applies wasn't established.
- The glossary says 429 responses carry Retry-After, while the main rate-limit section lists only the `X-Ratelimit` headers. We didn't trigger a 429 to confirm.
- No SLA was found for any plan in the terms or on the pricing page. A Premium contract may include one.
- No vendor MCP server for signing was found. The registry entry io.usefulapi/dropbox-sign is third-party and wasn't graded.

## Weaknesses

- No idempotency keys found in the docs or the spec, so a retried send can create a second signature request
- An API key grants full access to the account, with no scoped or read-only keys. Scopes exist only on OAuth tokens
- OAuth apps need manual approval by Dropbox Sign support before production use
- The terms supply the service as is, and no SLA was found. A major outage on 2 January 2026 lasted 3 hours 12 minutes
- Embedded signing and bulk send need the Standard plan ($3,000 a year), and embedded templates need Premium, priced by quote

## 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.

- Send `test_mode=true` while building. Test requests are free, watermarked and not legally binding, and are limited to 10 requests a minute
- Don't blind-retry a send after a timeout. No idempotency key exists, so list requests by `metadata` or title first to check whether it was created
- Authenticate with HTTP Basic, the API key as username and an empty password. Keep the key out of URLs, although the docs show that form
- Answer every webhook with HTTP 200 and the body `Hello API Event Received`, and verify `event_hash`. Ten consecutive failures clear the callback URL
- Treat a 200 from cancel as queued only. Confirmation arrives later as a `signature_request_canceled` event

## 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.