MCP Task Orchestrator
local MCP serverMCP Task Orchestrator is a local MCP server: Server-enforced workflow discipline for AI agents: work items, dependency graphs, quality gates. Install it from oci as ghcr.io/jpicklyk/task-orchestrator:3.15.0.
Installation
Terminal
claude mcp add io-github-jpicklyk-task-orchestrator -e DATABASE_PATH=... -e USE_FLYWAY=... -e AGENT_CONFIG_DIR=... -e MCP_TRANSPORT=... -e MCP_HTTP_PORT=... -e LOG_LEVEL=... -e LOG_FILE=... -- docker run -i --rm ghcr.io/jpicklyk/task-orchestrator:3.15.0
Run in your project folder. Add --scope user to use it everywhere.
Claude Desktop, Cursor & other clients
Add to claude_desktop_config.json or .cursor/mcp.json.
{
"mcpServers": {
"io-github-jpicklyk-task-orchestrator": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/jpicklyk/task-orchestrator:3.15.0"
],
"env": {
"DATABASE_PATH": "...",
"USE_FLYWAY": "...",
"AGENT_CONFIG_DIR": "...",
"MCP_TRANSPORT": "...",
"MCP_HTTP_PORT": "...",
"LOG_LEVEL": "...",
"LOG_FILE": "..."
}
}
}
}Fill in the environment variables with your own credentials.
What you need
- Runtime: Docker.
- MCP client: Claude Code, Codex, Antigravity, Claude Desktop, Cursor or any other client that speaks MCP over stdio.
- Environment variables:
DATABASE_PATH,USE_FLYWAY,AGENT_CONFIG_DIR,MCP_TRANSPORT,MCP_HTTP_PORT,LOG_LEVEL,LOG_FILE
Summary
Server-enforced workflow discipline for AI agents: work items, dependency graphs, quality gates
MCP Task Orchestrator vs similar servers
| Server | Stars | Type | API key | Runtime |
|---|---|---|---|---|
| MCP Task Orchestrator | 206 | Local | Yes | oci |
| socraticode | 3,322 | Local | Yes | npm |
| mcp-ssh-manager | 493 | Local | Yes | npm |
| google-analytics-mcp | 242 | Local | Yes | pypi |
| claude-prompts-mcp | 187 | Local | No | npm |
Questions
It reads 7 environment variables: DATABASE_PATH, USE_FLYWAY, AGENT_CONFIG_DIR, MCP_TRANSPORT, MCP_HTTP_PORT, LOG_LEVEL, LOG_FILE. Set them in your MCP config before starting the server.
Docker, plus an MCP client such as Claude Code, Codex, Claude Desktop or Cursor.
MCP Task Orchestrator is a local MCP server distributed as the oci package ghcr.io/jpicklyk/task-orchestrator:3.15.0.
Run: claude mcp add io-github-jpicklyk-task-orchestrator -e DATABASE_PATH=... -e USE_FLYWAY=... -e AGENT_CONFIG_DIR=... -e MCP_TRANSPORT=... -e MCP_HTTP_PORT=... -e LOG_LEVEL=... -e LOG_FILE=... -- docker run -i --rm ghcr.io/jpicklyk/task-orchestrator:3.15.0
Run: codex mcp add io-github-jpicklyk-task-orchestrator --env DATABASE_PATH=... --env USE_FLYWAY=... --env AGENT_CONFIG_DIR=... --env MCP_TRANSPORT=... --env MCP_HTTP_PORT=... --env LOG_LEVEL=... --env LOG_FILE=... -- docker run -i --rm ghcr.io/jpicklyk/task-orchestrator:3.15.0
Similar MCP servers
NameStars
- socraticodeio.github.giancarloerra/socraticode3,322
- mcp-ssh-managerio.github.bvisible/mcp-ssh-manager493
- google-analytics-mcpio.github.surendranb/google-analytics-mcp242
- claude-prompts-mcpio.github.minipuft/claude-prompts-mcp187
- wigoloio.github.KnockOutEZ/wigolo5,418
- DBHubio.github.bytebase/dbhub3,574
- Radar Kubernetes MCP Serverio.github.skyhook-io/radar3,517
- brightdata-mcpio.github.brightdata/brightdata-mcp2,660