# Fix list: Deno Sandbox From Anchor Terminal's listing at https://www.anchorterminal.com/tools/deno-sandbox, the October 2026 research run, assessed 8 October 2026. Grade D, 50.3 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 Deno Sandbox: 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. Reliability, 48 out of 100, up to 10.4 more on the total Why it scored 48: Hosted lines. denostatus.com runs on Instatus with history, but has no component for sandboxes (15 of 20). The last 90 days show one incident, increased request failures in `ord` on 7 September 2026, 1 hour 9 minutes to resolution with recovery reported after 11 minutes. Its effect on sandboxes is not stated (20). Concurrency and resource limits are published, though the docs (5 per organisation) and the pricing table (3 per region on Pro) differ, and no API request limits were found (8). No 429 or retry guidance and no idempotency keys found (0). A 99.95 per cent SLA is listed for Enterprise without saying whether it covers a beta product (5). Beta, pre-release per the docs (0). 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. ## 2. Payments & pricing, 20 out of 100, up to 10 more on the total Why it scored 20: No x402, MPP or L402 (0). Unit prices are public without a login, $0.10 a CPU-hour, $0.025 a GiB-hour and $0.20 a GiB-month (20). The Free plan does not include sandboxes, and no trial was found (0). A person signs up at console.deno.com in a browser and creates the token (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. ## 3. Security & auth, 61 out of 100, up to 6.8 more on the total Why it scored 61: One organisation token with no scopes or expiry found, rotatable from the dashboard, which also manages the organisation's Deploy apps. All members hold owner permissions (15). An outbound allowlist and read-only volume mounts exist, but outbound access is open unless `allowNet` is set, per the Security page (13). Secret substitution outside the VM is documented as a defence against prompt injection followed by exfiltration (13). Commands, HTTP requests and SSH sessions are traced in the dashboard with an event log (10). A disclosure policy with safe harbour at security@deno.com and a security.txt without an Expires field. The product page claims SOC2 and ISO27001 while the pricing table lists SOC2 Type 1 under Enterprise. No bug bounty found (10). `exposeHttp` URLs are public without authentication, which the docs state in capitals. 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. Agent ergonomics, 60 out of 100, up to 6.5 more on the total Why it scored 60: Sandbox and volume objects are small, and command output streams (15). Lists use cursor pagination with search, label and status filters (15). The Python SDK raises errors with a status, code, message and trace id, and validation errors with a field path, but the codes are not documented (10). No idempotency keys. A duration sandbox can be rejoined by id, and volume slugs are unique per organisation (5). `Sandbox.create()` needs no parameters, with JavaScript and Python SDKs and a CLI (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, 66 out of 100, up to 5.5 more on the total Why it scored 66: The published OpenAPI document has no sandbox operations, so the typed TypeScript SDK reference on JSR and the typed Python SDK count as 15 of 25 (15). llms.txt and every docs page as Markdown (10). The docs state use cases, when to choose a duration timeout and when to move work to a Deploy app (13). Typed options with stated ranges for memory, capacity and timeout (12). Examples in JavaScript, sync and async Python and the CLI. No error reference was found, and two pages disagree on the default network policy (8). SDKs are versioned at 0.x, and the Deploy changelog's newest entry is 12 March 2026 (8). 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. Maintenance & community, 45 out of 100, up to 4.8 more on the total Why it scored 45: `deno-sandbox` 0.16.0 on 22 July 2026, 78 days ago. `@deno/sandbox` last shipped 0.13.2 on 16 March 2026 (20). One release in the last 90 days (0). The Deploy changelog stops at 12 March 2026. Support is by Discord, and by email on paid plans. The Python repository shows one open issue or pull request (7). Official JavaScript and Python SDKs, the JavaScript one unchanged for over six months (12). The Python repository runs formatting, lint, type checks and offline tests in CI. The JavaScript SDK's source repository was not found in public (6). 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, 58 out of 100, up to 3.7 more on the total Made of editorial 35, provenance 80. Why it scored 58: The SDKs are MIT and the platform is closed under terms last modified 30 September 2026 (18). The privacy policy states no retention periods. The docs state that a deleted volume's storage is removed after 24 hours and the pricing table gives log retention of 1 day to 1 week. The terms grant Deno a licence over user data for internal business purposes including benchmarking. A DPA is listed for Enterprise only (10). No deprecation policy found, and the terms allow changes or discontinuation without notice (0). Two regions are named. No sub-processor list found (7). 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): - Endpoint on the vendor's domain: is not on deno.com (0 of 15) - Terms of service: read, states 6 of the 7 things a reader expects, and has 2 clauses that cost points (5.1 of 10) ## Deductions Each comes off the total. A fixed and documented problem counts for less at the next check. - 2026-10-08: the Create page (last modified 28 January 2026) says a default sandbox has no outbound network access, while the Security page of the same date says outbound access is unrestricted unless `allowNet` is set. A reader of the first page would leave egress open. 2 points. https://docs.deno.com/sandbox/create/ and https://docs.deno.com/sandbox/security/ ## 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 named a REST API for sandboxes. The docs link one, but its OpenAPI document has no sandbox, volume or snapshot operations. The Python SDK calls `/api/v3/sandboxes` paths that are not in any published contract we found. - Which default applies to outbound network access. The Create and Security pages disagree, and we did not run a sandbox to test it. - Whether the concurrency limit is 5 per organisation (docs) or 3 per region on Pro (pricing table). - Whether the Enterprise 99.95 per cent SLA covers sandboxes while they are in beta. - Which certification is current. The product page says SOC2 and ISO27001, the pricing table says SOC2 Type 1 under Enterprise, and no report or trust centre was found. - unchecked: the source repository and changelog for `@deno/sandbox`. github.com/denoland/sandbox asked for credentials, so it is private or absent. - unchecked: whether organisation tokens can be scoped or given an expiry in the console, which needs a login. - unchecked: issue response times on denoland/sandbox-py. Only the open count (1) was read. - No DPA or sub-processor list was found in public. deno.com/deploy/dpa returned 404. ## Weaknesses - Beta since 3 February 2026. The docs call the present phase pre-release, and no general availability date was found - Sandboxes are not included in the Free plan. Access starts at Pro, $20 a month, after a browser signup - The published OpenAPI document at `api.deno.com/v2/openapi.json` has 34 operations and none for sandboxes, volumes or snapshots - The Create page says a default sandbox has no outbound network access, and the Security page says outbound access is unrestricted by default - Maximum lifetime is 30 minutes, volumes exist only in `ord`, and no API rate limits or 429 guidance were found - Organisation tokens carry no scopes in the reviewed docs, and every organisation member has owner permissions ## 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. - Set `DENO_DEPLOY_TOKEN` to an organisation token (prefix `ddo_`) from Settings in console.deno.com. The organisation must be on Pro or above - Pass `allowNet` on every `Sandbox.create()`. The Security page says outbound access is unrestricted when it is omitted - Pass credentials through `secrets` with a `hosts` list, not `env`, so code in the VM sees only a placeholder - The default timeout ends the VM when the client disconnects. Pass a duration such as `"10m"` and reconnect with `Sandbox.connect({ id })`, up to 30 minutes - Create volumes in `ord` and start the sandbox in `ord`. A volume mounts only in its own region - `exposeHttp` URLs are public with no authentication. Treat the random subdomain as a secret ## 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.