"A task involving critical or creative thinking" matching MCP servers:

A secure MCP server providing controlled access to pixi tasks and commands, eliminating agent bash circumvention. It exposes 11 pixi tools through 4 meta-tools for ~95% context savings.

Enables asynchronous task execution and scheduling using RabbitMQ, Celery, and Redis/SQLite. Supports immediate background task execution, scheduled tasks with ISO 8601 timestamps, task status monitoring, cancellation, and completion notifications.

MCP server for publishing HTML or Markdown to a live hosted URL via htmldrop. Provides tools to publish, list, and delete hosted sites, with remote OAuth and API token authentication.

Provides MCP tools to get a single-call snapshot of git and GitHub Actions CI status across all repositories in a folder, flag missing or stale architecture docs, write doc files, and detect release drift between source and release repositories.

Enables AI coding tools to validate local projects and publish or update applications to Bedengan App Hub via API, with automated checks and live deployment without manual approval.

Detects breaking changes in a TypeScript library's public API and recommends the semver bump, comparing two npm releases, git refs or a working tree. Each finding is graded proven or review-only.

MCP server providing tools to verify task completion (tests pass and changes match claim) and confirm bug reproduction before fixing.

Record an agent's MCP tool-call workflow once, replay it deterministically for zero tokens, and get a receipt — a snapshot-testing and deterministic-execution layer for Agent Skills.

Stops your AI from re-introducing bugs, leaking provider keys, or weakening tests. Bug fixes become permanent regression guards; blocked mistakes become AI lessons the agent reads and learns from before its next edit.

Enables AI coding agents to compute a 0-1 diff risk score for the current branch or any pull request, exposing the driving signals, evidence, tier and resulting auto-approve or human-review policy decision. Agents can use it to decide whether a low-blast-radius, AI-reviewed PR is safe to merge automatically while flagging risky changes for a human.

Enables code-review gating by converting a git diff into a typed verdict with safety probabilities, severity, and recommended actions for auto-merge or human review.

Deploys and manages projects on Swigs Cloud from any MCP client, turning folders or git repositories into live HTTPS sites in about two minutes.

MCP server exposing Apache Airflow REST API operations as tools — list DAGs, inspect runs and task instances, trigger DAG runs, and check failed DAGs and scheduler health

Publishes what an AI made — a page, a markdown or PDF doc, or a whole site — to a live HTTPS link. Runs locally over stdio from the @drop2run/mcp npm package, or hosted at https://mcp.dropto.run/mcp for chats that connect by URL.

The network-change verifier for AI agents: validate a BGP or routing change in a real containerlab lab before it reaches production, and get a pass, fail or inconclusive verdict with evidence. Runs on FRR, Nokia SR Linux, SONiC and Arista cEOS.

An MCP server that enables AI assistants to interact with Apache Airflow's REST API for DAG management, task monitoring, and system diagnostics. It provides comprehensive tools for triggering workflows, retrieving logs, and inspecting system health across Airflow 2.x and 3.x versions.

Runs a project's Playwright or CTRF test suite and returns a redacted failure packet (failing test, file and line, console errors, failed requests, rerun command) so a coding agent can fix and rerun. Also reports whether a commit has a signed, passing test receipt on GitHub.

Blocks unsafe PostgreSQL migrations before an AI agent writes or runs them. check_before_apply returns a pass/fail gate; reads the real Postgres parser, classifies the lock each statement takes, checks 112 safety rules. Runs offline, no database required.

Lets any MCP-capable client connect to the public Codemagic REST API to list apps and teams, inspect build history, and optionally trigger or cancel builds and manage team settings. Runs locally over stdio with the user's own API token and configurable read-only, build, or administrative permission modes.

Enables MCP clients to read a Forgejo, Codeberg, or Gitea instance—including users, repositories, issues, pull requests, search, notifications, comments, reviews, and Actions runs—with opt-in, time-boxed guarded writes for repository, branch, issue, pull request, comment, migration, push-mirror, and workflow operations.

Checks whether software dependencies (npm packages, GitHub Actions, MCP servers, Docker images) are alive, dormant, abandoned, archived, or deleted using dated observations from lastseen.dev.

Enables inspection and management of Apache Airflow DAGs, runs, and logs across multiple instances. It provides tools for monitoring workflows and performing gated write operations like triggering DAGs or clearing task instances.

Post-quantum readiness for AI coding agents: scans code for quantum-vulnerable cryptography (RSA/ECDH/ECDSA/DH) and returns NIST ML-KEM/ML-DSA/SLH-DSA (and hybrid) migration guidance, with fix verification and dependency checks. Content-based, advisory tools only; runs local (npx @quantakrypto/mcp) or as a hosted OAuth endpoint.

Enables agents and CI pipelines to audit MCP servers and agent tool-chains by statically scanning repositories, local checkouts, tools/list exports, or live endpoints for risks such as destructive actions without confirmation, mismatched safety annotations, injection surfaces, credential or PII exposure, and unguarded command, path, or URL sinks. All checks are read-only and never execute the scanned code or call tools/call.

Read-only Playwright CI results for coding agents: tells whether a failing test is new in this PR, already failing on main, or known flaky, with per-test history.

Enables LLMs to manage mainframe source code and deployment workflows via BMC Compuware ISPW, supporting assignment, task, release, and deployment operations.

A comprehensive and efficient Model Context Protocol server for task management that works with Claude, Cursor, and other MCP clients, providing powerful search, filtering, and organization capabilities across multiple file formats.

Enables management of Railway.app infrastructure through natural language, including deploying services from GitHub repos or Docker images, managing environment variables, monitoring deployments, and configuring networking and volumes.

Lets MCP clients such as Claude or ChatGPT find the pull requests waiting on your review, read their details and changed files, check CI status, post comments, and submit approve, request-changes, or comment reviews. It runs locally over stdio with a personal token or as a remote server with its own OAuth 2.1 authorization server (PKCE, dynamic client registration, per-client consent, and refresh-token rotation) so GitHub is only used to sign the user in.

An MCP server that automates the full software development lifecycle through an AI-driven TDD state machine. It handles everything from task decomposition and test-driven development to integration testing and automated pull request creation.