lynxprompt-mcp
According to the provided schema, this MCP server lets an LLM search LynxPrompt’s public blueprint marketplace and create, update, and delete blueprints and hierarchies.
Search public blueprints by query, with optional tags, type, and category filters.
Create blueprints with name, type, content, and optional visibility/description; PUBLIC visibility publishes them.
Update existing blueprints by ID, optionally changing name, content, visibility, or description.
Delete blueprints by ID.
Create hierarchies with name and repository_root, plus optional description.
Delete hierarchies by ID.
Schema note: all six tools are annotated as non-read-only, open-world, non-idempotent, and destructive, including search_blueprints; the README also describes additional read-only tools/resources not present in this schema.
It works with lynxprompt.geiser.cloud or your own self-hosted LynxPrompt, over HTTP or stdio.
Features
Tools to search the public marketplace, list your own blueprints and hierarchies, read any blueprint you can see, and create, update and delete them. Every tool carries a title, read-only/destructive hints and an output schema.
update_blueprinttakes thecontent_checksumfromget_blueprintasexpected_checksum, so it never silently overwrites an edit made elsewhere.LYNXPROMPT_READ_ONLY=trueregisters only the read tools.Read-only resources too:
lynxprompt://blueprints,lynxprompt://blueprint/{id},lynxprompt://hierarchies,lynxprompt://hierarchy/{id},lynxprompt://user.Speaks MCP 2026-07-28 and every earlier revision down to 2024-11-05 from the same binary, so current clients (which still use the
initializehandshake) and 2026-07-28 clients both work.stdio, or stateless Streamable HTTP on
/mcpwith Origin checks, a body limit and bearer auth when exposed.Ships as an npm package (
npx lynxprompt-mcp), a Docker image and multi-arch Go binaries.
Related MCP server: mcp-openapi-schema
Tools
Tool | What it does | Changes data |
| Search the public marketplace (query, configs/commands, category, tags, sort, page) | no |
| Your blueprints, paged, filter by visibility | no |
| Any blueprint you can see (yours, your team's, public) with content and | no |
| Your hierarchies (monorepo groupings), paged | no |
| One hierarchy with its blueprints, flat and as a tree | no |
| New blueprint; | yes |
| Change fields; pass | yes, overwrites |
| Permanent delete | yes, destructive |
| New hierarchy; returns the existing one for the same | yes |
| Delete a hierarchy; its blueprints are kept, unlinked | yes, destructive |
Install
Create an API token on your LynxPrompt instance under Settings, API tokens. For a self-hosted instance also set LYNXPROMPT_URL. Add LYNXPROMPT_READ_ONLY=true to hide the write tools.
Claude Code
claude mcp add lynxprompt --scope user --env LYNXPROMPT_TOKEN=lp_... -- npx -y lynxprompt-mcpClaude Desktop (claude_desktop_config.json) and Cursor (~/.cursor/mcp.json)
{
"mcpServers": {
"lynxprompt": {
"command": "npx",
"args": ["-y", "lynxprompt-mcp"],
"env": { "LYNXPROMPT_TOKEN": "lp_..." }
}
}
}VS Code (.vscode/mcp.json, asks for the token once and stores it encrypted)
{
"inputs": [
{ "type": "promptString", "id": "lynxprompt-token", "description": "LynxPrompt API token", "password": true }
],
"servers": {
"lynxprompt": {
"type": "stdio",
"command": "npx",
"args": ["-y", "lynxprompt-mcp"],
"env": { "LYNXPROMPT_TOKEN": "${input:lynxprompt-token}" }
}
}
}Codex (~/.codex/config.toml)
[mcp_servers.lynxprompt]
command = "npx"
args = ["-y", "lynxprompt-mcp"]
env = { LYNXPROMPT_TOKEN = "lp_..." }Docker (stdio, any client): command docker, args ["run", "-i", "--rm", "-e", "TRANSPORT=stdio", "-e", "LYNXPROMPT_TOKEN", "drumsergio/lynxprompt-mcp:v0.2.0"].
The npm package downloads the release binary on first run and checks its sha256 against the checksums shipped inside the package. The HTTP server, Docker Compose and local builds are in Getting started.
Documentation
Getting started: Docker Compose, npm, local build
Configuration: environment variables and MCP client configuration
Usage: resources, tools and protocol versions
Development: testing, contributing, credits
Related projects: the family and registry listings
Related projects
LynxPrompt, the self-hosted platform this server talks to.
License
Available Tools
10 toolscreate_blueprintCreate blueprintA
Save a new blueprint (an AI agent instruction file such as AGENTS.md or CLAUDE.md) to the token user's LynxPrompt account. Each call creates a new blueprint, even with the same name; to change an existing one use update_blueprint. Private by default; visibility PUBLIC publishes it to the public marketplace. To place it in a monorepo hierarchy pass hierarchy_id and repository_path. Returns the new blueprint with its bp_ id.
Inputs: name (required), content (required), tags, type, +7 more. Returns: success, blueprint
create_hierarchyCreate hierarchyAIdempotent
Create a hierarchy: a group that ties the blueprints of one repository together, for example every AGENTS.md of a monorepo. repository_root identifies the repository; if you already have a hierarchy with that repository_root, it is returned unchanged instead of a duplicate. Add blueprints to it with create_blueprint (hierarchy_id plus repository_path). Returns the hierarchy with its ha_ id.
Inputs: name (required), repository_root (required), description. Returns: tree, message, success, hierarchy, +2 more
delete_blueprintDelete blueprintADestructiveIdempotent
Permanently delete one of your blueprints. It cannot be restored. Confirm the id with get_blueprint or list_blueprints first.
Inputs: id (required). Returns: message, success
delete_hierarchyDelete hierarchyADestructiveIdempotent
Permanently delete one of your hierarchies. The blueprints in it are not deleted: they stay in your account, unlinked from any hierarchy.
Inputs: id (required). Returns: message, success
get_blueprintGet blueprintARead-onlyIdempotent
Read one blueprint the token can see, including its full content and its content_checksum: your own, your team's, or a public one found with search_blueprints. Call it before update_blueprint and pass content_checksum as expected_checksum, so an edit made elsewhere in the meantime is not overwritten. Only the owner can update or delete a blueprint. For a public blueprint of someone outside your team, repository_path is only the file name.
Inputs: id (required). Returns: success, blueprint
get_hierarchyGet my hierarchyARead-onlyIdempotent
Read one of your hierarchies with every blueprint in it, both as a flat list (blueprints[], with repository_path and content_checksum) and as a tree nested by parent_id. Blueprint content is not included; use get_blueprint for that.
Inputs: id (required). Returns: tree, message, success, hierarchy, +2 more
list_blueprintsList my blueprintsARead-onlyIdempotent
List the blueprints owned by the user of LYNXPROMPT_TOKEN, newest change first, without their content. Use it to find a blueprint id (bp_...) for get_blueprint, update_blueprint or delete_blueprint. Paged: when has_more is true, call again with offset increased by limit.
Inputs: limit, offset, visibility. Returns: limit, total, offset, has_more, +1 more
list_hierarchiesList my hierarchiesARead-onlyIdempotent
List your hierarchies. A hierarchy groups the blueprints of one repository (for example the AGENTS.md files of a monorepo) under one repository_root. Returns ids (ha_...) for get_hierarchy and delete_hierarchy. Paged like list_blueprints.
Inputs: limit, offset. Returns: limit, total, offset, has_more, +1 more
search_blueprintsSearch marketplace blueprintsARead-onlyIdempotent
Search the public LynxPrompt marketplace of AI agent instruction files (AGENTS.md, CLAUDE.md, Cursor rules, slash commands and so on). Use it to find blueprints other people published; get_blueprint reads a result's content. To find your own blueprints use list_blueprints instead. Needs no token. Returns one page of results (templates[] with id, name, description, author, type, tags) plus total and hasMore.
Inputs: page, sort, tags, type, +3 more. Returns: total, hasMore, templates, popularTags
update_blueprintUpdate blueprintADestructiveIdempotent
Change one of your blueprints. Only the fields you pass change; content replaces the whole text and the old text is not kept. Read the blueprint with get_blueprint first and pass its content_checksum as expected_checksum: if someone edited it since (web UI, CLI, another agent) the call fails with 409 instead of overwriting their change. Setting visibility PUBLIC publishes it to the marketplace.
Inputs: id (required), name, tags, type, +6 more. Returns: success, blueprint
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 10 tools
Every tool maps to a distinct resource (blueprint vs hierarchy) plus a distinct action, and the descriptions explicitly resolve the two potentially confusable pairs: list_blueprints (own) vs search_blueprints (public marketplace), and create_blueprint (always new) vs update_blueprint (modify existing). An agent can pick the right tool without guessing.
Perfect verb_noun snake_case pattern throughout: create_/get_/list_/update_/delete_/search_ combined with _blueprint or _hierarchy. The two nouns mirror the two resources, so the naming itself reinforces the mental model.
10 tools for a two-resource domain (blueprints and hierarchies) with full lifecycle plus marketplace search is well-scoped; each tool earns its place and none is redundant.
Blueprints have full coverage: create, get, list, search, update, delete, with optimistic-concurrency via content_checksum. Hierarchies cover create/get/list/delete but have no update operation (no rename, and create_hierarchy returns the existing one unchanged on duplicate repository_root), and there is no explicit way to unlink a blueprint from a hierarchy without deleting it.
Maintenance
Related MCP Connectors
An MCP server that provides an API to LLMs to manage their JumpCloud resources.
MCP server providing access to the Scorecard API to evaluate and optimize LLM systems.
MCP server for generating rough-draft project plans from natural-language prompts.
Related MCP Servers
- AlicenseBqualityDmaintenanceAn MCP server that provides LLMs access to other LLMs456 npm83MIT
- AlicenseNot gradedqualityDmaintenanceA MCP server that exposes OpenAPI schema information to LLMs like Claude. This server allows an LLM to explore and understand large OpenAPI schemas through a set of specialized tools, without needing to load the whole schema into the context50MIT
- FlicenseNot gradedqualityDmaintenanceMCP server for PromptHub. Enables publishing, fetching, listing, updating, and searching prompt repositories from Claude Code or Codex.-
- FlicenseAqualityDmaintenanceAn MCP server that provides AI clients with access to developer documentation via llms.txt files. Exposes tools, resources, and prompts.31-