mcp-install
Install, validate, and manage MCP servers for OMK using the curated catalog and parallel processing.
git clone --depth 1 https://github.com/dmae97/open-multi-agent-kit /tmp/mcp-install && cp -r /tmp/mcp-install/templates/skills/kimi/mcp-install ~/.claude/skills/mcp-installSKILL.md
# MCP Server Installation Skill
Guide for adding MCP servers to an OMK project using the built-in catalog and parallel orchestration.
## Commands
```bash
# List configured MCP servers
omk mcp list
# Diagnose all MCP servers
omk mcp doctor
# Diagnose a specific server
omk mcp doctor <server-name>
# Install a single MCP server
omk mcp install <name> <command> [args...]
# Remove a server from project-local config
omk mcp remove <name>
# Sync global MCP servers into project
omk mcp sync-global
```
## Curated Catalog
OMK ships with a verified catalog in `src/mcp/server-catalog.ts` (or accessible via the runtime).
Recommended categories:
- **reasoning**: `sequential-thinking` — structured problem solving
- **memory**: `memory` — persistent knowledge graph
- **docs**: `pdf` — PDF text extraction
- **web**: `puppeteer` — browser automation
- **devtools**: `filesystem`, `sqlite` — file and database access
- **ops**: `railway` — remote Railway MCP; `supabase` — database management and edge functions
## Parallel Installation
When installing multiple servers (e.g., during `omk init`), OMK uses `Promise.all` for parallel setup.
Agents and hooks can also trigger bulk installs:
1. **Agent lane**: explorer validates which servers are needed
2. **Coder lane**: writes the MCP JSON entries
3. **QA lane**: runs `omk mcp doctor` to verify each server
4. **Hook**: `post-init-mcp.sh` runs non-blocking validation after init
## Rules
1. Prefer `npx -y <package>` so no global install is required.
2. Use `${PROJECT_ROOT}` and `${DB_PATH}` placeholders in catalog entries; they are resolved at install time.
3. Do not commit secrets in `mcp.json` env values.
4. Run `omk mcp doctor` after adding servers to confirm they start.
5. If a server fails to start, check `startup_timeout_sec` and network access.
## When to Use
- During `omk init` when selecting additional MCP servers.
- When a task requires a new capability (browser, PDF, database).
- Before reporting an MCP issue, run `omk mcp doctor --json` for structured evidence.Persistent memory, recall, session replay, and memory-governance workflow adapted from rohitg00/agentmemory for OMK. Use when setting up agent memory, deciding what to remember, importing/replaying sessions, reducing repeated context, or auditing memory safety.
Minimal, goal-driven, surgical coding workflow adapted from forrestchang/andrej-karpathy-skills for OMK. Use for coding, refactoring, debugging, and review tasks where assumptions, overengineering, or broad edits could cause regressions.
Legal workflow drafting, triage, review, research planning, legal operations, law-student or clinic support, and legal AI governance adapted from Anthropic claude-for-legal. Use for commercial, privacy, product, corporate, employment, regulatory, AI governance, IP, litigation, legal-clinic, and law-student tasks. Draft-only; attorney review and current source verification required.
Real-engineering alignment, shared-language, TDD, diagnosis, and architecture-review workflow adapted from mattpocock/skills for OMK. Use before non-trivial implementation, ambiguous product work, debugging loops, test-first changes, or codebase architecture cleanup.
Managed-agent teamwork, issue assignment, progress tracking, reusable-skill compounding, and handoff workflow adapted from multica-ai/multica for OMK. Use when coordinating multiple agents, converting work into agent-ready tasks, tracking blockers, or turning repeated solutions into skills.
Review AdaptOrch, OMK, and similar DAG multi-agent orchestration frameworks. Use when assessing DAG node responsibility, dependency edges, worker write authority, fallback/retry/timeout/evidence gates, review/merge boundaries, or reproducible decision traces.
Optional read-only OMK web/social/video research workflow inspired by Panniantong/Agent-Reach. Use for web search, current social evidence, YouTube/Bilibili/Reddit/Twitter/X/RSS/GitHub public research, and Agent Reach availability checks without auto-installing or collecting credentials.
Backend API review for NestJS, Express, FastAPI, database access, validation, auth, error handling, and API contracts.