ScienceLogo MCP server
Commit Nov 9, 2026Enables AI agents to write, maintain, and validate ScienceLogo workflows and libraries through a structural reference reader.
Also known as ScienceLogo/MCP, mcp
- Stars
- 0
- Tools
- 24
- Downloads, 30 days
- —
- Latest version
- 0.2.0
- Last commit
- Nov 9, 2026
- License
- MIT
What is ScienceLogo MCP?
ScienceLogo enables AI agents to write, maintain, and validate ScienceLogo workflows and libraries through a structural reference reader. It exposes 24 tools, such as check, outline, read_workflow.
Summary generated from the project's README
About ScienceLogo MCP server
Managing scientific workflows and libraries through AI agents is possible with ScienceLogo. This MCP server allows users to write, maintain, and validate structured workflows using a reference reader that checks every write for structural correctness. If an agent attempts an invalid change, the reader rejects it and provides the specific file, line, and column of the error.
Developers can use this server with MCP clients like Claude Code, Codex, or Goose via stdio transport. The setup requires the package to be installed locally with Racket and the ScienceLogo package available on your PATH. The server includes built-in resources like a guide and manual pages to teach the AI the specific grammar and syntax of the language.
Specific tools enable deep interaction with the language structure. You can use create_workflow to start new projects, write_fragment to add text, or convert_syntax to switch between verbose and compact modes. The server also provides specialized commands like declare_procedure to define scientific steps with units and uncertainty, or add_paired_stage to manage entry and exit states.
What you can do with it
- 1Validate workflow syntaxAsk the AI to check a ScienceLogo file for structural errors and readability advice.
- 2Define scientific proceduresInstruct the agent to create a new procedure including its inputs, results, and uncertainty levels.
- 3Explore existing librariesHave the AI describe the procedures and record interfaces available in a library before importing it.
- 4Convert file syntaxRequest the agent to rewrite a workflow file from verbose to compact syntax while preserving comments.
Written from the project's README. Every use case is checked against a quote from it.
How to install ScienceLogo MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall ScienceLogo MCP in Claude Code
$ claude mcp add mcp -- docker run -i --rm ghcr.io/sciencelogo/sciencelogo-mcp:0.2.0Install ScienceLogo MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"mcp": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/sciencelogo/sciencelogo-mcp:0.2.0"
]
}
}
}Install ScienceLogo MCP in VS Code
.vscode/mcp.json
{
"servers": {
"mcp": {
"type": "stdio",
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/sciencelogo/sciencelogo-mcp:0.2.0"
]
}
}
}Install ScienceLogo MCP in VS Code CLI
$ code --add-mcp '{"name":"mcp","type":"stdio","command":"docker","args":["run","-i","--rm","ghcr.io/sciencelogo/sciencelogo-mcp:0.2.0"]}'Install ScienceLogo MCP in Claude Desktop
claude_desktop_config.json
{
"mcpServers": {
"mcp": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/sciencelogo/sciencelogo-mcp:0.2.0"
]
}
}
}Install ScienceLogo MCP in Codex
$ codex mcp add mcp -- docker run -i --rm ghcr.io/sciencelogo/sciencelogo-mcp:0.2.0Install ScienceLogo MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.mcp] command = "docker" args = ["run", "-i", "--rm", "ghcr.io/sciencelogo/sciencelogo-mcp:0.2.0"]
Install ScienceLogo MCP in Gemini CLI
$ gemini mcp add -s user mcp docker -- run -i --rm ghcr.io/sciencelogo/sciencelogo-mcp:0.2.0Install ScienceLogo MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"mcp": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/sciencelogo/sciencelogo-mcp:0.2.0"
]
}
}
}ScienceLogo MCP tools 24
Detected from the READMEcheckoutlineread_workflowlanguage_referencedescribe_librarycreate_workflowcreate_librarywrite_fragmentremove_itemmove_itemconvert_syntaxset_workflow
| check | — | |
| outline | — | |
| read_workflow | — | |
| language_reference | — | |
| describe_library | — | |
| create_workflow | — | |
| create_library | — | |
| write_fragment | — | |
| remove_item | — | |
| move_item | — | |
| convert_syntax | — | |
| set_workflow | — |
Show all 24 toolsShow fewer tools
declare_performerdeclare_resourcedeclare_propositiondeclare_interfacedeclare_procedureadd_stageadd_paired_stageadd_phaseadd_requirementadd_mappingimport_libraryadd_include
| declare_performer | — | |
| declare_resource | — | |
| declare_proposition | — | |
| declare_interface | — | |
| declare_procedure | — | |
| add_stage | — | |
| add_paired_stage | — | |
| add_phase | — | |
| add_requirement | — | |
| add_mapping | — | |
| import_library | — | |
| add_include | — |
Automated checks
Ran Nov 14, 2026Full report →Not verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachablesciencelogo/mcp is public and reachablegithub.com/sciencelogo/mcp ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)MITspdx.org/licenses/MIT.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-11-09github.com/sciencelogo/mcp ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model24 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubScienceLogo MCP server
An MCP server that lets AI agents, and through them people, write and maintain ScienceLogo workflows and libraries. It works the same way in Claude Code, Codex, Goose and other MCP clients.
The ScienceLogo reference reader checks every write before it is saved. A write the reader rejects changes nothing, and the agent gets the reader's message with the file, line and column.
Tools
Items are named kind:name, for example stage:Measure. Every tool that
writes is checked by the reader, and a construct tool replaces an item of the
same kind and name.
Reading and checking
| Tool | What it does |
|---|---|
check |
Runs the reader, the structural assessment and readability advice. |
outline |
Lists every item, with the stages, phases, procedures, performers and imports. |
read_workflow |
Returns a workflow in verbose or compact syntax, whole or one item; a file the current reader rejects is returned as stored. |
language_reference |
Returns a manual page, the grammar, the changelog or an example of the installed release. |
describe_library |
Shows what a library provides, its procedures and record interfaces, before it is imported. |
Starting and changing the structure
| Tool | What it does |
|---|---|
create_workflow |
Starts a workflow with its title and summary. |
create_library |
Starts a library project. |
write_fragment |
Writes any ScienceLogo text, at the end of a file or next to or in place of an item. |
remove_item |
Removes an item. |
move_item |
Moves an item to another file, keeping the method's order. |
convert_syntax |
Rewrites files in verbose or compact syntax, keeping comments. |
Writing each construct
| Tool | What it writes |
|---|---|
set_workflow |
The workflow's title and summary. |
declare_performer |
Who or what carries out work. |
declare_resource |
A resource the method uses, such as an instrument. |
declare_proposition |
A hypothesis, assumption or conclusion. |
declare_interface |
A record interface: the fields a record must or may have. |
declare_procedure |
A procedure, with its result's fields and units and its uncertainty. |
add_stage |
A simple stage, in its own file under stages/ by default. |
add_paired_stage |
A stage with entry and exit state; the exit state is derived from the calls. |
add_phase |
A named group of stages with its own summary. |
add_requirement |
A required call or stage order. |
add_mapping |
A statement that a record implements an interface. |
import_library |
A library import pinned to a tag or commit. |
add_include |
An include of an existing file. |
Resources: a guide for agents (sciencelogo://guide) and the manual
pages (sciencelogo://docs/{topic}). No model knows ScienceLogo, so the
server teaches it.
Steps done outside ScienceLogo are described as a method, never by file
paths or code: declare_procedure asks for what the step does, its inputs,
the kind, fields and units of its result, and its uncertainty.
New stages get their own file under stages/, and the advice returned with
each write helps keep the workflow readable from the broad description down
to the details.
MCP access
The server is published with this identity:
- registry name:
io.github.ScienceLogo/sciencelogo-mcp - OCI image:
ghcr.io/sciencelogo/sciencelogo-mcp:<release-version>, with Racket and ScienceLogo inside - transport:
stdio - official registry:
https://registry.modelcontextprotocol.io/
It is discoverable in the official MCP Registry under
io.github.ScienceLogo/sciencelogo-mcp:
https://registry.modelcontextprotocol.io/?q=sciencelogo
Local client configuration, with the package installed and Racket with the
sciencelogo package on the PATH (or named by SCIENCELOGO_MCP_RACKET):
{
"mcpServers": {
"sciencelogo": {
"command": "sciencelogo-mcp"
}
}
}
The registry metadata is declared in server.json. The
publish workflow fills in the release version and image tag when publishing.
Versions
The first two numbers of the server's version name the ScienceLogo language
version it supports: 0.2.x works with language 0.2. See
CHANGELOG.md.
Development
scripts/validate-local.sh
runs ruff, builds the image and runs the tests inside it, against the
image's Racket and ScienceLogo. Set CONTAINER=podman to use Podman.
Pushing a tag vX.Y.Z builds the image, pushes it to the GitHub Container
Registry and publishes server.json to the MCP Registry.
Licence and citation
MIT, copyright 2026 Riccardo Boero. See LICENSE.
If you use the ScienceLogo MCP server in published work, please cite it as:
Boero, R. (2026). ScienceLogo MCP Server [Software]. Zenodo. https://doi.org/10.5281/zenodo.24085876
ScienceLogo MCP questions
How do I add ScienceLogo MCP to Claude Code?
Run claude mcp add mcp -- docker run -i --rm ghcr.io/sciencelogo/sciencelogo-mcp:0.2.0 in your terminal. Add --scope user to make it available in every project.
How do I add ScienceLogo MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
What tools does ScienceLogo MCP provide?
It exposes 24 tools, including check, outline, read_workflow, language_reference, describe_library.
How does the server ensure workflow validity?
The ScienceLogo reference reader checks every write before it is saved, rejecting invalid changes and providing the agent with the specific file, line, and column of the error.
What happens when a tool attempts to overwrite an existing item?
A construct tool replaces an item of the same kind and name.