Hexis Cloud

# Fix list: Hexis Cloud

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/hexis, the October 2026 research run, assessed 14 October 2026. Grade C, 55.3 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 Hexis Cloud: 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, 18 out of 100, up to 16.4 more on the total

Why it scored 18: Graded on the hosted lines, because the surface is Hexis Cloud's MCP endpoint. No status page found, and status.hexis.software answers 'No workspace at this address' (0). No readable incident history (5). The security page says sign-in attempts are rate limited, but we found no request limits for the MCP endpoint or any tool (0). No documented 429, Retry-After or backoff guidance (0). The terms state no uptime figure and no SLA unless an order form has one, and the pricing section lists SLAs only for Enterprise, with no terms published (3 of 10). The service carries no beta label, though it went live on 11 October 2026 and the software is version 0.29.0 (10). The terms also say there are no backups and the whole service runs on one server, which this checklist doesn't score but a reader should weigh.

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, 30 out of 100, up to 8.8 more on the total

Why it scored 30: No x402, MPP or L402 (0). Plan pricing is public, Free for up to 3 seats and Team at €10 a seat a month excluding VAT, with no per-call unit (10). The Free plan has no time limit and asks for no card (20). A person must sign in with Google, Microsoft or a password and tick the terms to create a workspace. OAuth dynamic client registration lets an MCP client register itself, but a person still completes the sign-in (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. Agent ergonomics, 57 out of 100, up to 7 more on the total

Why it scored 57: About 37 platform tools plus three code-mode tools are listed directly, before any connected tool, which puts the listing over 30 (5). Code mode (list_tools, tools_info and call_tool_chain in a sandbox), skill lookup through list_skills and get_skill, and connection-key callers seeing only tools whose credentials are set add back 7 (12 of 25). grep and list_files narrow results, and call_tool_chain takes max_output_size and spills a large result to a file read back with read_file (14 of 20). Refusals name the problem, a missing branch is answered 404 naming it, and list_tool_setup returns setup.reason with the next step (15 of 20). No readOnlyHint or destructiveHint annotations. Writes land on a branch, and protected branches change only through approved change requests, which limits the harm of a repeated call (8 of 20). Branch defaults on read tools. One official client, the TypeScript local server on npm (8 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.

## 4. Security & auth, 64 out of 100, up to 6.3 more on the total

Why it scored 64: OAuth 2.1 with PKCE, dynamic client registration and token revocation, but a single `mcp` scope, and connection keys stored as SHA-256 hashes and revocable. The agent always acts with the connecting person's permissions (25 of 30). Read access is denied by default, rules apply per folder and file, a connection can be read-only, and protected content changes only with its owners' approval (17 of 20). The tools return knowledge-base pages any writer can edit and the results of third-party tools, and we found no prompt-injection guidance (3 of 15). Admins see an audit log of agent tool calls, who and with which connection, kept 365 days by default, and can revoke any connection (14 of 15). A valid security.txt and a candid security page, but the same page says there is no certification, no penetration test, no automated vulnerability scanning and no bug bounty, and the encryption keys sit on the server with the data (5 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.

## 5. Schema & documentation, 83 out of 100, up to 2.8 more on the total

Why it scored 83: Every tool on tools/list has a JSON Schema input, and the server validates each schema before listing it so one malformed tool can't blank the list (25). No llms.txt. The agent guide is Markdown in the repository and goes to agents in the MCP handshake instructions, with the first line repeated on the four core tools (7 of 10). Descriptions open with a call example, say what each verdict such as canWrite means and when to open a change request instead, and point to the guide's shared rules (15 of 20). Inputs carry required fields, minLength, integer bounds and additionalProperties false, with some free-form code strings for call_tool_chain (12 of 15). Call examples on every tool and named refusals such as branch-required, but no single error reference (11 of 15). Release notes for every version on GitHub, on 0.x versions with no stated versioning policy (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.

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

Made of editorial 81, provenance 69.

Why it scored 75: Hexis is Apache-2.0 and the terms describe Hexis Cloud as a managed version of it. Releases name an extension point for distributions, so we can't confirm every Cloud-only piece is public (27 of 30). The Hexis Cloud privacy policy, DPA, security page and data-portability page agree on retention (logs capped at 150 MB, deleted workspaces kept 30 days, audit log 365 days) and say no AI provider and no analytics service receives data. Bevel's own privacy policy at bevel.software/privacy describes a different product with OpenAI and Serper as processors and carries two different dates, which a reader may confuse with Hexis Cloud's (24 of 30). The terms give six weeks' notice of price changes and a right to leave on service changes, with no deprecation policy for tools or endpoints (8 of 20). Sub-processors are listed with locations and transfer safeguards, and a change log (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):

- Domain age: hexis.software, no registry record we could read (0 of 15)
- Terms of service: published, but our reader couldn't read it (7 of 10)
- Privacy policy: published, but our reader couldn't read it (7 of 10)
- Status page: not found (0 of 10)

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

Why it scored 85: v0.29.0 on 14 October 2026 (30). 36 version tags since v0.1.0 on 10 August 2026, eight of them in October (20). No issues have ever been opened on GitHub, and the open items are eight pull requests, mostly from the maintainers, reviewed with comments the same week. Little outside activity to judge responsiveness by (15 of 25). In the official MCP registry as io.github.Bevel-Software/hexis, but the entry is version 0.1.0 from 11 August and points at the demo instance, not Hexis Cloud (10 of 15). A test workflow on every push and pull request with a real Postgres, a lint ratchet and dependencies pinned to exact versions (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 Product Hunt launch date of 12 October. The legal documents date Hexis Cloud to 11 October 2026, and we didn't read Product Hunt
- unchecked: the tool list a live Hexis Cloud workspace returns. We read it from the source and created no workspace
- unchecked: whether Hexis Cloud runs any code that isn't in the public repository
- Bevel's site describes '100s of tools' and '58 ready-made skills'. We didn't count either. Tools are any MCP server or API a workspace connects
- The MCP registry entry still points at demo.bevel.software at version 0.1.0

## Weaknesses

- The terms and security page state no backups, no SLA, no penetration test, no certification and no MFA of its own
- All workspaces share one Hetzner server in Nuremberg with logical separation, and the encryption keys sit on that server
- No status page and no published request rate limits for the MCP endpoint
- Over 35 platform tools are listed directly, with no readOnlyHint or destructiveHint annotations
- Hexis Cloud is open to businesses only, and a person must sign in with Google, Microsoft or a password to create a workspace

## 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 person for a read-only connection when the task only reads. It answers canWrite false everywhere
- Pass the branch you are working on to every knowledge-base tool. Calls without a required branch are refused
- Check canWrite from list_files, read_file or get_skill before writing, and open a change request when it is false
- Use list_tools and tools_info before call_tool_chain, and never ask the person for a tool's secret in chat
- Keep your own copy of anything important. Hexis Cloud keeps no backups of Managed repositories

## 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.