"Servers Offering MCP Prompts as Prompt Primitives" matching MCP servers:

An MCP server that allows agents to test and compare LLM prompts across OpenAI and Anthropic models, supporting single tests, side-by-side comparisons, and multi-turn conversations.

Enables users to manage, search, render, version, and regression-test local Markdown templates with typed variables, and to expose them as MCP tools and native prompts without requiring a model API, account, or cloud service.

A local MCP server for prompt evaluation, enabling users to define test cases, run prompts against multiple LLM providers, score outputs with deterministic and LLM-graded assertions, and track quality over time, all within an AI coding environment.

Enables prompt optimization loops and regression test suites for Claude Code, with a companion web UI for real-time visualization of scores and prompt revisions.

Enables LLMs to inspect and test other MCP servers by listing and calling their tools, reading resources, and getting prompts. Supports stdio, SSE, and HTTP transports for both local and remote MCP server inspection.

A validation layer for AI coding assistants that enforces explicit LLM evaluations on plans, code diffs, and tests to ensure safer and higher-quality code.

Provides prompt templates and tools for mobile QA automation, including converting manual tests to Appium scripts, analyzing failures, and standardizing bug reports.

Enables running deterministic multi-model prompt regression and golden dataset benchmark suites, exposing structured results and telemetry through the Model Context Protocol for integration with MCP-compliant clients.

A remote MCP server exposing 3 tools via Model Context Protocol: score_prompt (free, grades any LLM prompt A-F on 0-40 scale), optimize_prompt (returns optimized prompt + dimension breakdown, $0.025 USDC), and compare_models (Claude vs GPT-4o head-to-head, $1.25 USDC). Remote HTTP server at /api/mcp.

Enables Claude to build and preview mobile apps on multiple iOS and Android devices simultaneously, with live streaming and touch control, all via MCP.

MCP server for the Ghost Inspector API that lets agents inspect and manage end-to-end browser tests, offering read-only tools for inventory, module usage, and stale test analysis, plus opt-in guarded write operations.

Free MCP tools for context engineering audits of AI agents, offering on-demand analysis of system prompts, tool definitions, and context packing.

Enables AI agents to hire real human operators for tasks requiring physical presence, human perception, or judgment, such as verification, testing, data collection, and physical-world tasks.

MCP server for AI security scanning. Detects prompt injections, jailbreaks, data exfiltration, and social engineering attacks in real-time using 42 regex patterns via the ClawGuard Shield API. Sub-10ms response times.

Enables AI coding tools to run CheckMyLaunch scans on GitHub repos or live URLs and retrieve findings and repair prompts directly within the chat.

A thin MCP server that wraps provenance-receipts and quality-gate services as tools, enabling AI agents to certify content origin and score quality via Ed25519-signed receipts.

Enables AI agents to hire verified human operators for tasks requiring physical presence, human perception, or judgment, such as real-world verification, product testing, and data collection.

Renders HTML in a headless browser and detects layout defects numerically, such as text overlaps, placeholder leftovers, and element overflow, without requiring baseline images.

Exposes MAID Runner validation tools (validate, snapshot, test) as MCP tools, resources, and prompts for AI agents like Claude Code and Aider, enabling seamless integration with AI development workflows.

Security scanner for AI agent packages that enables AI agents to audit MCP servers and packages for vulnerabilities, prompt injection, and supply chain attacks.

Enables AI coding agents and IDEs to prove SystemVerilog assertions with SymbiYosys (smtbmc+z3) and lint SVA syntax, returning honest verdicts such as PROVEN, FAILED, UNKNOWN, ERROR, or TIMEOUT.

An MCP server that grades prompts against a validated rubric, tracks real outcomes, and provides guidance grounded in research on what predicts good results.

Enables LLM agents to programmatically inspect, debug, and test other MCP servers by wrapping the MCP Inspector CLI. Supports listing and calling tools, reading resources, and testing prompts on both local and remote MCP servers.

Exposes Semgrep as deterministic, rule-matching analysis tools over MCP so a code-review agent can scan only the lines changed in a diff, scan whole files on demand, and check analyzer availability, version, ruleset, and limits. Results are returned as compact per-finding lines with source snippets suitable as comment anchors, along with counts of suppressed out-of-diff matches.

Checks a website or local dev server with axe-core (WCAG 2.0/2.1 A and AA) and returns a fix prompt with real examples from the pages, for the AI tool to fix the source. Automated checks find roughly a third of accessibility problems.

Render, analyze, and verify audio (WAV or FLAC) through a fully offline, deterministic engine, exposed as MCP tools for AI agents.

MCP server exposing SuperAudit’s free website audit as a callable tool, enabling AI agents to audit any public URL and receive scores across SEO, security, legal, performance, and more. No API key required; respects rate limits and caches results.

Provides coding agents with visibility into test health through tools for flaky test detection, test quality linting, and LLM evaluation harness, enabling them to triage failures, review test quality, and check prompt changes for regressions.

Enables MCP clients to score retrieval quality with recall, precision, nDCG, hit rate and MRR, compare 2-3 configurations with seeded paired-bootstrap confidence intervals, audit agent tool calls against expected traces, and detect prompt regressions against golden sets. It also generates Markdown, JSON or JUnit reports with pass/fail gates, all deterministically and offline without paid API keys.

Checks that every computed number in a Markdown document still matches the formula that produced it, and repairs it under an explicit write gate — so an agent's correct formula can't ship with a wrong total. Exposes every VisiMark command as a tool (check, explain, eval, infer, format), the authoring discipline as resources, and the orderings an agent gets wrong unsupplied as prompts. Read-only by