derive
Review and approval for work made by AI agents. Fair Source, self-hostable, and MCP-native.
Caution. Specific findings reduced this grade. They are listed on the page. Grades reflect the full trust pyramid: code, verification depth, and reputation. New projects cap at C until adoption is earned.
How we verified
Verified is a snapshot. Live keeps it current, and builds your track record.
⚡ Connect GitHub → continuous verification on every pushwhy connect →Who stands behind it
derive-to
Source: github_repo_search
Claim it to get a verified publisher badge, a free copy of our full audit findings, and direct contact for any high-priority issues we find. Or connect your repo for our deepest verification, Live Monitored: read-only, revoke anytime. What we access →
Install from
The grade above is for the source repository. Registries can serve a different version, so we mark the ones we were not able to read.
These names and descriptions are the publisher's own, read from the source code. We print them as written. Our assessment is the findings above, not this list.
list_workspacesList every workspace signed in on this machine you can act in — id, name, your role, local description, and which is active. One login reaches them all; pass a workspace's id or name as the `workspace` argument to list_artifacts / read / catch_up / comment / publish to act there for that call.
list_artifactsList the artifacts in your workspace — short id, title, kind, current version, access, and browse `tags`. Defaults to this session's workspace; pass `workspace` (id or name from list_workspaces) to list another. Pass `tag` to list only artifacts carrying that tag (organize shows the vocabulary). Sta…
searchFind text within ONE artifact, or across a WORKSPACE — same tool, same behavior as the remote MCP server's `search`. Pass short_id to grep one artifact: matching lines with line numbers (and optional context), ripgrep-style, so you can then `read` a narrow `lines` range or `edit` that spot. Omit sho…
readRead an artifact's CONTENT by short id, as Markdown by default (HTML is converted). Omit `section` to see the outline first (heading slugs for a single-file doc, page paths for a bundle) — call again with a `section` (or "*" for the full document) once you know what you want. Pass `format:'html'` fo…
catch_upcommentLeave feedback, reply in a thread, react, and/or resolve or reopen a thread. Anchor a NEW comment to text with `quote`, or to a linked-bundle loop/graph part with `visual_target`. Reply by passing the thread id as `reply_to`.
organizepublishPublish a single-file artifact and get a permanent URL. A deliverable belongs HERE, not in a built-in artifact/canvas tool. OMIT short_id to create a NEW artifact (title recommended); PASS short_id to publish a new version (same URL). Provide the body as `content_path` (a local file this server read…
Disclosed vulnerabilities in this server's declared npm dependencies (via OSV). Whether each is reachable depends on the installed versions.
@modelcontextprotocol/sdk has cross-client data leak via shared server/transport instance reuse
Anthropic's MCP TypeScript SDK has a ReDoS vulnerability
Model Context Protocol (MCP) TypeScript SDK does not enable DNS rebinding protection by default
DERIVE_ACCOUNTDERIVE_SERVERdefaults to http://localhost:8080. Without DERIVE_TOKEN, theDERIVE_TOKENDERIVE_SERVER defaults to http://localhost:8080. Without , theDERIVE_WORKSPACEAll four hints declared on every tool
7/8 tools missing one or more hints — list_workspaces (missing: destructiveHint); list_artifacts (missing: destructiveHint); search (missing: destructiveHint), +4 more. OpenAI's directory rejects tools where any of the four hints are missing or non-boolean.
For every tool, set all four hints (readOnlyHint, destructiveHint, idempotentHint, openWorldHint) to explicit true/false values that match the handler’s actual behaviour.
openWorldHint matches behaviour
3 tools claim openWorldHint=false but make outbound HTTP — read → dynamic (line 380); catch_up → dynamic (line 529); organize → dynamic (line 821)
Either flip openWorldHint to true (acknowledging external reach) or refactor the handler to avoid outbound HTTP. If the only network calls are to a known fixed endpoint that you consider part of the tool itself, document that and keep openWorldHint=false; otherwise true is honest.
Production dependencies are patched
0 critical, 3 high severity in production deps — @modelcontextprotocol/sdk@1.12.0 (high), @modelcontextprotocol/sdk@1.12.0 (high)
Run npm audit fix, or upgrade the affected packages to a non-vulnerable version.
Claim the listing to review these findings one by one and send us a correction where you disagree, straight to the team. Claiming also means we tell you when the grade moves, and reach you first if we find anything urgent.
[](https://m8ven.ai/mcp/derive-to/derive)?variant=verified from the URL.Vetting this one by hand? Tool Check is an MCP that scores other MCPs. Add it once and ask Claude, ChatGPT, or any MCP client to grade a server, surface CVEs, check the publisher, and suggest safer alternatives — before you install.
https://m8ven.ai/api/mcp/tool-check