guilds.run

# Fix list: guilds.run

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/guilds-run, the October 2026 research run, assessed 14 October 2026. Grade C, 56.5 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 guilds.run: 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, 53 out of 100, up to 9.4 more on the total

Why it scored 53: Read with the local-software lines, because the owner runs it. Docker with Compose is the stated requirement, and the install is a git clone of main plus ./setup.sh, which builds the images. No package, tagged release or published image exists, since the image workflow runs only on a v* tag and there are none (10 of 20). The verify workflow runs typecheck, lint, orchestrator, UI and web tests and a smoke test of the Compose stack, and all 13 runs on main from 9 to 13 October 2026 passed. The clone has 196 test files (25). No open issues, and the three issues filed were closed within an hour, on five days of public history (18 of 25). No version numbers, tags or changelog (0). Not versioned or declared stable (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. Schema & documentation, 58 out of 100, up to 6.8 more on the total

Why it scored 58: Framework reading. No OpenAPI document for the HTTP API. The recipe format guilds.recipe v2 rejects unknown keys, and connector manifests are validated with TypeBox at load (12 of 25). No llms.txt on guilds.run or docs.guilds.run (404). The docs are Markdown in the repository with an AGENTS.md (5 of 10). Pages state what the code does today, which page to read for which job, and the trust boundary in numbered rules (15 of 20). Tool input schemas, recipe caps and typed manifest fields (11 of 15). The getting-started walk says what you should see at each step, and the Operate page explains each startup error message (11 of 15). No changelog or versioned releases. The security register is the one page that keeps dated history (4 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.

## 3. Security & auth, 68 out of 100, up to 5.6 more on the total

Why it scored 68: Framework reading. 30 for what leaves the machine by default, 20 for approvals and sandboxing, 15 for prompt injection, 15 for audit and 20 for the security programme. Model keys go only to their provider, Sentry starts only when SENTRY_DSN is set with user info, bodies and model inputs and outputs excluded, and Graft runs with DO_NOT_TRACK=1. Graft's once-a-day npm version lookup has no switch (27). Each machine is an OpenSandbox container with dropped capabilities and no_new_privileges, writing MCP tools start Off and Studio changes wait for the operator. Agents on one machine share it and run commands as root with open egress (SEC-6, SEC-24), no per-call approval gate for tool calls was found, and the default box has no login (10). The coding prompt says repository content is data, Studio wraps evidence as untrusted, and known secret values are redacted from transcripts. SEC-8, injection into coding agents, is open (8). Every run keeps the prompt, each tool call and result, and versioned memory with who changed it (14). SECURITY.md with GitHub private reporting and a public register of 26 issues with exposure and residual risk, 13 still open. No security.txt, bounty or published advisory (9).

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, 67 out of 100, up to 5.4 more on the total

Why it scored 67: Framework reading, scored on what it takes to run a tool-calling team with MCP. No code is needed. Three commands and a model key start a guild, and each agent is offered only the tools turned on for it (18 of 25). max_tool_calls defaults to 50 per run, max_input_tokens trims old tool results and hands off to a new run, and remaining_tool_calls is appended to each result after the platform cap (15 of 20). Runs end with named codes such as generation.max_tool_calls and workspace.full, and interrupted tool calls get an error result so the next model request is valid (13 of 20). readOnlyHint and destructiveHint are stored for each MCP tool and writing tools start Off. Startup does not replay a model call (14 of 20). Sensible defaults and a guided first-run dialog, but it is an application with an HTTP API and no SDK in any language (7 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.

## 5. Maintenance & community, 41 out of 100, up to 5.2 more on the total

Why it scored 41: No tagged release. Main, which the install clones and SECURITY.md names as the supported version, last changed on 13 October 2026, so 20 of 30 rather than full marks for a release. No releases or dated changelog entries (0). Eight issues and pull requests, all closed, the three issues closed within an hour, one with two comments, on five days of history (15 of 25). No published package (0). CI on every push with a frozen pnpm lockfile, no Dependabot configuration in the repository (6 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.

## 6. Payments & pricing, 60 out of 100, up to 5 more on the total

Why it scored 60: No payment protocol (0). Self-hosted rule. The software is free under AGPL-3.0 with nothing to buy today (20), needs no card (20) and needs no account with Cigale AI (20). The hosted edition at $19 a month is listed as coming soon and was not scored.

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.

## 7. Transparency & trust, 52 out of 100, up to 4.2 more on the total

Made of editorial 63, provenance 40.

Why it scored 52: AGPL-3.0-only, OSI approved (30). No privacy policy or terms. The footer links to both and they return 404. The docs say where data lives (PostgreSQL and STATE_ROOT on your host), that model keys go only to their provider, and that stored credentials are encrypted under one master key in .env (12 of 30). No deprecation policy or notices. The security register is dated (3 of 20). Sentry is opt-in and documented on the Configure page, and the Graft telemetry setting and its version lookup are disclosed on the code tools page (18 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):

- Domain age: guilds.run, registered 2026-09-19 (under a year) (0 of 15)
- Status page: not found (0 of 10)
- Changelog: not found (0 of 10)
- security.txt: not found (0 of 10)

## Deductions

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

- 2026-10-14. The home page says each agent has its own computer, and its script example reads sandboxed in scout's workspace. Since the change of 13 October 2026 (pull request #8, per-org machines), agents are placed on shared machines, and the vendor's own register (SEC-24, open and accepted) says agents on one machine can read each other's homes, git tokens, browser sessions and running commands' environments. The footer also links to terms and privacy pages that return 404. The docs state the shared-machine risk plainly, so 2 points. https://www.guilds.run/ and https://github.com/CigaleAI/guilds/blob/main/docs/architecture/security-register.md

## 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's Product Hunt date of 12 October 2026 is not the first release. The public repository was created on 9 October 2026 (GitHub API). The founder's earlier availability announcement was not found first-hand, and the @guildsrun posts on X were not read
- unchecked: the Product Hunt page, which answered with a Cloudflare bot check
- Whether a legal form and address exist for Cigale AI. None was found on the site or in the repository
- Whether the terms and privacy pages linked from the footer are planned. Both returned 404 on 14 October 2026
- The home page lists GLM under subscriptions, and the core providers table names Kimi Code as the only coding provider
- The home page says each agent has its own computer. Since 13 October agents share machines per organisation, per pull request #8 and SEC-24

## Weaknesses

- No tagged release, version number, changelog or published image. Users install whatever is on main
- 13 issues in the vendor's security register are open, among them shared machines (SEC-24), root shells (SEC-6) and prompt injection into coding agents (SEC-8)
- No login on the default box. Whoever reaches port 8080 is the operator, and remote use needs a proxy or tailnet
- Terms and privacy links on guilds.run return 404, and no privacy statement for the software was found
- Five days of public history on 14 October 2026, with 57 stars and eight issues and pull requests

## 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.

- Keep the box on loopback or behind an authenticating proxy with AUTH_MODE=trusted-header. The default mode does not authenticate
- Put agents that hold different secrets or GitHub tokens on different machines. Agents on one machine can read each other's homes and tokens
- Back up the .env file written by ./setup.sh. It holds the only copy of the key that encrypts stored credentials
- Set a spend limit at your model provider. guilds.run sets no spend cap, and max_tool_calls defaults to 50 per run
- Pin a commit of main when you deploy. There are no tagged releases or versioned images yet

## 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.