Skip to main content
Glama
tanmays0

OpsMCP

by tanmays0
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Local

OpsMCP

MCP (Model Context Protocol) server that exposes eight typed tools so an AI agent can operate GitHub, Postgres, deploy status, and a sandboxed local filesystem. Credentials stay in the server process; tool inputs are treated as untrusted.

Repository: https://github.com/tanmays0/ops-mcp
Control plane: https://tanmays0.github.io/ops-mcp/

Artifact

Location

MCP server (stdio)

uv run python -m ops_mcp

Eight tools

registered in src/ops_mcp/server.py

Companion UI

site/ → GitHub Pages (catalog, setup, simulated playground)

Seeded demo database

docker compose up -d → Postgres users / orders / products

Automated tests

uv run pytest (path sandbox, SQL guard, GitHub mocks, Postgres)

CI

GitHub Actions workflow .github/workflows/ci.yml

Pages deploy

.github/workflows/pages.yml

Container image

Dockerfile

Agent wiring example

.cursor/mcp.json.example

Architecture

Agent (stdio) → FastMCP server
                  ├── tools/       domain tools
                  ├── security/    path sandbox, SQL allowlist, redaction, rate limit
                  ├── adapters/    GitHub HTTP, Postgres
                  └── config.py    Pydantic Settings (env / .env)

Related MCP server: GithubMCP

Tools

Tool

System

Behavior

github_list_issues

GitHub API

List issues/PRs by repo, state, labels

github_create_issue

GitHub API

Create issue; default dry_run=true

github_pr_diff_summary

GitHub API

PR file list and diff stats (no patches)

postgres_query_readonly

Postgres

Parameterized SELECT only

postgres_explain

Postgres

EXPLAIN (FORMAT TEXT) only (no ANALYZE)

deploy_status

GitHub Actions

Latest workflow run for a branch

fs_search

Local FS

Literal search under allowlisted roots

fs_read_file

Local FS

Read file with sandbox and size limit

Safety

Control

Implementation

Filesystem

Resolve-then-compare allowlist after symlink resolution

SQL

sqlglot AST: single SELECT / WITH…SELECT; DDL/DML/multi-statement rejected before connect

GitHub writes

dry_run=true unless explicitly disabled

Secrets

SecretStr for token and DSN; log redaction; never returned in tool payloads

Outbound HTTP

Token-bucket rate limit

Stack

Python 3.12 · FastMCP · httpx · psycopg · sqlglot · Pydantic Settings · pytest · Docker

Run

git clone https://github.com/tanmays0/ops-mcp.git
cd ops-mcp
uv sync
docker compose up -d
cp .env.example .env
uv run pytest
uv run python -m ops_mcp

Populate .env from .env.example (OPS_MCP_FS_ROOTS, OPS_MCP_DATABASE_URL, OPS_MCP_GITHUB_TOKEN). MCP client config: copy .cursor/mcp.json.example to .cursor/mcp.json with absolute paths. Secrets load from .env only (.env and .cursor/mcp.json are gitignored).

Configuration

Variable

Purpose

OPS_MCP_FS_ROOTS

Colon-separated allowlisted directories

OPS_MCP_FS_MAX_READ_BYTES

Max bytes for fs_read_file

OPS_MCP_DATABASE_URL

Postgres DSN

OPS_MCP_GITHUB_TOKEN

GitHub fine-grained or classic PAT

OPS_MCP_HTTP_RATE_PER_SECOND

Outbound API rate

OPS_MCP_LOG_LEVEL

Log level

Verification flows

Documented end-to-end flows: DEMO.md.

Spec

Project principles: .specify/memory/constitution.md
Requirements: specs/001-ops-mcp-server/spec.md

This server cannot be deployed

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • enterprise-agent-lab

    isaac-dev-ai
    F
    license
    Not graded
    quality
    C
    maintenance
    Enables controlled AI-agent access to enterprise-shaped tools with a deny-by-default gated write path, human approval, dry-run execution, and append-only audit logging.
    1
    -

  • GithubMCP

    Dhananjayrbiraris
    A
    license
    A
    quality
    B
    maintenance
    Connects AI assistants to GitHub repositories, pull requests, issues, commits, and code search while enabling repository visibility controls, CI/CD monitoring, sandboxed local filesystem access, and code quality/security analysis.
    13
    1
    MIT

  • MCP ToolHub

    zhihaochen67
    F
    license
    A
    quality
    A
    maintenance
    Enables coding agents to perform workspace-confined file operations, read-only Git inspection, and structured shell commands, while requiring out-of-band human approval for mutations and external executions and maintaining an audit trail.
    14
    3
    -

  • Sandbox MCP

    Dhananjay-JSR
    A
    license
    A
    quality
    B
    maintenance
    Enables AI coding agents to safely execute commands, run tests, and modify project files inside disposable, policy-enforced Docker sandboxes that are isolated from the host machine and its credentials.
    15
    MIT