CCM
/MCP
SkillsMCPMarketplacesDigestToolsAdvertise

This week in Claude

Every Monday: Claude Code, Agent SDK, MCP, and the Anthropic platform moves worth your time.

Skills by Category
Frontend DevelopmentBackend & APIsTesting & QASecurityDevOps & CI/CDGit & Pull RequestsDocumentationCode Review & QualityAI & Agent BuildingSkill Development
MCP Servers by Category
Sales & MarketingWeb & Browser AutomationDatabasesAI & LLM ToolsCloud & InfrastructureCommunication & MessagingDeveloper ToolsDesign & CreativeDocuments & KnowledgeSearch & Web Crawling
Marketplaces by Category
AI Agents & OrchestrationLLM IntegrationDevelopment ToolsFrontend & UIBackend & APIsDatabasesTesting & Code QualityDevOps & CloudSecurity & ComplianceGit & Version Control

Claude Code Marketplaces

Discover Claude Code plugins, extensions, and tools. Automatically updated directory of Anthropic Claude AI marketplaces with development tools, productivity plugins, and integrations.

Resources

  • Browse Skills
  • Browse MCP Servers
  • Browse Marketplaces
  • Skill index
  • MCP index
  • Marketplace index
  • Plugins Reference

Community

  • About
  • Tools
  • Feedback
  • Privacy Policy
  • Advertise

Built for the Claude Code community with Claude Code by mertbuilds.com

Independent project, not affiliated with Anthropic
seojeom avatar

Seojeom Mcp

seojeom/seojeom-mcp
STDIOregistry active
Summary

Binds Claude to the project currently open in the Seojeom desktop app through a shared registry directory, then exposes local wiki and graph operations over stdio. You get read, search, and write tools for wiki pages, plus graph query and mutation primitives with approval gates on destructive operations. Runs in router mode by default, which means one MCP connection serves whichever project is active rather than hardcoding to a single path. Install via npx from the public package or run from source for the full authoring surface including prompts and resources. Reach for this if you're already using Seojeom for project knowledge graphs and want Claude to read and update them directly.

CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
MCP-ready Email SendingMCP-ready Email Sending
MCP-ready Email Sending
Plug Mailtrap into your AI workflow and let it handle the email.
Connect Mailtrap MCP →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Capacitor - Shared memory for your team’s coding agents.
Capacitor - Shared memory for your team’s coding agents.
Make coding agent sessions - Searchable, Shareable, Vendor-neutral & Scored.
Try For Free →
CodeScene MCP ServerCodeScene MCP Server
CodeScene MCP Server
Your agent targets a perfect 10 Code Health score. Deterministic. Every commit.
Try For Free →
Give your AI the whole web as clean markdownGive your AI the whole web as clean markdown
Give your AI the whole web as clean markdown
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
inference shell
inference shell
create and run specialised agents in minutes
build now →
CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
MCP-ready Email SendingMCP-ready Email Sending
MCP-ready Email Sending
Plug Mailtrap into your AI workflow and let it handle the email.
Connect Mailtrap MCP →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Capacitor - Shared memory for your team’s coding agents.
Capacitor - Shared memory for your team’s coding agents.
Make coding agent sessions - Searchable, Shareable, Vendor-neutral & Scored.
Try For Free →
CodeScene MCP ServerCodeScene MCP Server
CodeScene MCP Server
Your agent targets a perfect 10 Code Health score. Deterministic. Every commit.
Try For Free →
Give your AI the whole web as clean markdownGive your AI the whole web as clean markdown
Give your AI the whole web as clean markdown
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
inference shell
inference shell
create and run specialised agents in minutes
build now →

Seojeom MCP Server

Canonical repo location: packages/seojeom-mcp

During the Godot-only decommission transition, apps/desktop/mcp-server may already be absent. New work should treat this package path as the source of truth. If a deprecated local lane still needs the old path to exist, recreate the minimal metadata-wrapper with pnpm apply:legacy-compat-wrapper.

Local-first MCP server for Claude and Codex.

seojeom-mcp binds to the project currently opened by the Seojeom desktop app or Godot launcher through a shared registry directory, then serves that project's local wiki and graph data over stdio.

Quickstart

Claude Code

claude mcp add -s local \
  -e SEOJEOM_SHARED_REGISTRY_DIR=/mnt/c/Users/<you>/.seojeom/registry \
  seojeom -- npm exec --yes --package=seojeom-mcp seojeom-mcp -- \
  --router \
  --host-kind auto \
  --approval-mode prompt

Claude Desktop

Add this to claude_desktop_config.json and restart Claude Desktop:

{
  "mcpServers": {
    "seojeom": {
      "command": "npm",
      "env": {
        "SEOJEOM_SHARED_REGISTRY_DIR": "<shared-registry-dir>"
      },
      "args": [
        "exec",
        "--yes",
        "--package=seojeom-mcp",
        "seojeom-mcp",
        "--",
        "--router",
        "--host-kind", "auto",
        "--approval-mode", "prompt"
      ]
    }
  }
}

Print the exact onboarding command first

npm exec --yes --package=seojeom-mcp seojeom-mcp -- --print-claude-onboarding

What it provides

  • project binding to the currently active desktop/Godot project
  • local wiki read/search/write tools
  • local graph read/search/query tools
  • core graph mutation/proposal tools for local authoring tasks
  • approval-gated mutating operations

The exact tool surface is discovered at runtime through MCP tools/list.

Runtime contract

  • primary transport: stdio
  • primary install path: npm exec --yes --package=seojeom-mcp seojeom-mcp --
  • recommended mode: shared-registry router mode
  • expected runtime: Node >=20.20.0
  • public npm package shape: standalone stdio entrypoint only

Source build contract

Running pnpm build in this package is a dual-output build:

  • dist/index.js and the rest of dist/** stay available as the local source/runtime HTTP sidecar surface used by Godot and other repo-local lanes
  • .public-package/dist/standalone.js is the staged npm publish surface

After a successful build, you can verify both lanes with:

pnpm smoke:build-layout

Godot-side source verification can also run the local HTTP sidecar smoke directly from the canonical package root:

pnpm smoke:sidecar-health

Surface tiers

seojeom-mcp is not one uniform surface. Treat these as separate contracts:

1. npm public package

  • audience: external Claude/Codex users installing from npm
  • transport: stdio
  • mode: shared-registry router
  • published shape: LICENSE, README.md, bin/seojeom-mcp, dist/standalone.js, package.json
  • capability promise: public tools only
  • non-goals: internal prompts, internal resources, full authoring harness, desktop UI shell

This is the only contract guaranteed by the published npm package.

2. local full runtime

  • audience: local maintainers running from source
  • transport: stdio plus local sidecar HTTP runtime paths
  • profile: full authoring surface
  • includes: prompts, resources, specialized graph/wiki authoring flows, internal playbooks

This source-tree runtime is intentionally larger than the public package.

3. desktop-coupled runtime

  • audience: local desktop/Godot-integrated operation
  • product default: app-bundled sidecar runtime
  • depends on: shared registry writer, approval queue UI, review shell, sidecar health surface
  • examples: approval review pane, desktop approval alerts, graph IPC bearer-token flows
  • debug override: sidecar entry override is reserved for debug-only startup paths

These desktop-coupled capabilities are not part of the standalone npm guarantee.

Registry

  • MCP Registry identifier target: io.github.seojeom/seojeom-mcp
  • planned publish mode: GitHub Actions Trusted Publishing + npm publish --provenance

Notes

  • This package is still a preview release and the tool surface may evolve between versions.
  • Router mode is the recommended public path. Direct per-project standalone flags remain as a compatibility fallback.
  • The public package defaults to a core authoring tool surface. Prompts, resources, and more specialized orchestration surfaces are reserved for non-public/internal profiles.
  • Public package documentation intentionally stays focused on installation and end-user usage. Internal release automation and repository-specific operator workflows are not part of the package contract.

Links

  • Repository: https://github.com/seojeom/seojeom-mcp
  • Issues: https://github.com/seojeom/seojeom-mcp/issues
  • License: MIT
Featured
CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
MCP-ready Email SendingMCP-ready Email Sending
MCP-ready Email Sending
Plug Mailtrap into your AI workflow and let it handle the email.
Connect Mailtrap MCP →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Capacitor - Shared memory for your team’s coding agents.
Capacitor - Shared memory for your team’s coding agents.
Make coding agent sessions - Searchable, Shareable, Vendor-neutral & Scored.
Try For Free →
CodeScene MCP ServerCodeScene MCP Server
CodeScene MCP Server
Your agent targets a perfect 10 Code Health score. Deterministic. Every commit.
Try For Free →
Give your AI the whole web as clean markdownGive your AI the whole web as clean markdown
Give your AI the whole web as clean markdown
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
inference shell
inference shell
create and run specialised agents in minutes
build now →
Categories
Documents & Knowledge
Registryactive
Package@seojeom/mcp-server
TransportSTDIO
UpdatedApr 23, 2026
View on GitHub

Related Documents & Knowledge MCP Servers

View all →
seritalien avatar
Brain Protocol

seritalien/brain-protocol

Verifiable Memory for AI Agents — knowledge graph, FTS, Starknet proofs, local or cloud
sketchjar avatar
OpenWarrant — AI Text Detection

sh.stipple/openwarrant-aitext

Detect AI-generated text: the probability a document's prose was LLM-written, with the tells.
sketchjar avatar
OpenWarrant — Reference Verification

sh.stipple/openwarrant-verify

Verify a document's citations resolve and match, recheck its math, flag unsupported claims.
sidharth-5u15 avatar
CupShup Marketing Knowledge Base

sidharth-5u15/cupshup-mcp

Search CupShup's marketing knowledge base — services, case studies, and insights from India's leading integrated marketing agency. Covering BTL, RWA activations, digital marketing, and brand activations across 500+ brands served since 2014. Every response links back to cupshup.co.in.
sidharth-5u15 avatar
CupShup Marketing Knowledge Base

sidharth-5u15/cupshup-mcp-6cb1f167

Search CupShup's marketing knowledge base — services, case studies, and insights from India's leading integrated marketing agency. Covering BTL, RWA activations, digital marketing, and brand activations across 500+ brands served since 2014. Every response links back to cupshup.co.in.
sketchjar avatar
OpenWarrant

sketchjar/openwarrant

Verify document authenticity for AI agents: detect tampering in PDFs and images with evidence.