"Information about draw.io diagramming tool" matching MCP servers:

Audits MCP tool descriptions for quality and reliability, scoring them 0-100, detecting smells, and providing rewritten versions for better agent accuracy.

Protects AI agents from bad tool outputs, schema mismatches, and tool poisoning by validating, sanitizing, and scoring reliability of tool responses.

Enables verification of deployment-scoped model-visible DSH tool surface conformance by checking recorded tool identities and schemas against approved locks, reporting only drift categories and hashes.

Analyzes multi-step AI agent tool chains to compute success probability, identify bottlenecks, and suggest better execution orders, enabling more reliable agents via local pure-math computation.

A lightweight multi-provider MCP testing fixture with two providers (eva_air and uk_railway), each exposing 5 tools using in-memory fake data, designed for verifying MCP client tool discovery, tool calls, and multi-provider routing.

This MCP server verifies that an agent’s claimed tool output matches the actual response returned by a tool to prevent invented or distorted results. It flags mismatches, omissions, and invented fields to ensure downstream logic only processes verified tool data.

MITM proxy MCP server that intercepts and verifies AI agent tool calls, detecting fabrication, tracking costs, verifying outcomes, and inferring user satisfaction.

A throwaway HTTP MCP server that empirically tests whether Cursor invokes tools by exact name or munges them, to inform case normalization decisions.

An MCP Server that provides access to Google's Cloud Tool Results API, enabling applications to programmatically interact with test results from Firebase Test Lab and mobile app testing through natural language.

Enables autonomous agents to automatically repair and guard tool-call schemas through deterministic AST analysis and parameter projection validation, ensuring reliable, schema-compliant interactions in MCP-compatible environments.

Exposes a fictional smart-home store's REST API to agents as schema-validated MCP tools, resources, and prompts, letting them search products, check inventory, place atomic orders, look up customers and orders, and open support tickets. It also adds auth injection, rate limiting, REST-to-MCP error mapping with actionable hints, per-call JSONL tracing, and a golden-task eval harness for tool selection and argument correctness.

An MCP server that exposes a document store for evaluating agent tool-use accuracy and measuring resistance to indirect prompt injection, with built-in conformance to the 2026-07-28 MCP protocol.

Enables researchers to present 12 mock file, memory, schedule, and database tools over stdio and collect LLM pre-execution judgments (EXECUTE/CONFIRM/REFUSE) without executing real operations.

An interactive tool that enables users to benchmark vLLM endpoints through MCP, allowing performance testing of LLM models with customizable parameters.

Audits local package folders to generate publish-readiness reports and previews of NPM tarball contents. It helps developers identify and resolve configuration issues like missing types or broken exports before releasing code.

Enables agents and operators to audit MCP tool permissions for least privilege, score blast radius, detect privilege-escalation chains, verify capability matrices, and produce cryptographic audit proofs with injection fuzzing and exfiltration canaries.

Enables AI agents to drive a local asset review loop: list and fetch paired original/output items (including image content for visual comparison), mark pass/reject/unsure with problem tags and region annotations, batch-review scoped selections, and export results as JSON or CSV. Also supports round archiving, detecting stale results after source re-renders, and listing previous-round rejects.

An MCP server that reads app source code, generates contextually valid dummy data, and drives a browser through pages with narrated steps, error correction, and output reports.

MCP server exposing statistical regression testing for LLM agents as a "run" tool: p-value, effect size, and confidence interval on whether agent behavior actually changed.

MCP server exposing a "run" tool that benchmarks agent-memory backends and returns a reproducible JSON report.

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

Enables recurring website maintenance checks — domain registration, SSL certificates, uptime, technical SEO and accessibility — across a portfolio of client sites, using publicly available information. Results are returned as structured data and can be sorted by urgency into a client-ready report.

Provides an in-memory notebook interface with tools to create, retrieve, list, update, and delete notes, designed to test agent behavior through realistic MCP tool interactions.

Enables coding agents to compile supplied files, run optional focused tests, and get static findings across ten programming languages, providing a deterministic feedback loop on the code they are about to return.

Static security linter for MCP servers. Scans tool definitions for vulnerabilities (path traversal, SQL injection, SSRF), scores description quality, and auto-rewrites descriptions for safer agent tool selection.

Enables AI assistants to summarize Playwright test runs, identify flaky tests, heuristically triage failures, generate Playwright test stubs, and retrieve recent GitHub CI status through MCP tool calls.

Exposes six evidence-oriented tools over stdio so MCP clients can navigate unfamiliar code, ask typed questions about repository evidence, review completed changes, and select existing tests.

Umbra is the trust layer for AI-generated code. It scores any repo 0-100 with file:line evidence (security, slop, Docker-verified build/boot, and claim receipts that catch agents lying about tests), and guards agent file writes mid-stream blocking leaked keys, alg:none JWTs, and git-hook plants before they land. Tools: scan_repo, guard_content, get_score.

Grades MCP servers on agent usability (A–F) — description quality, schema design, tool naming and token cost, the properties that decide whether a model picks the right tool. Benchmarked across 36 popular servers; a third scored D or F.

Open behavioral litmus for MCP servers — grades A–F across tool-output injection, egress, sensitive-data, and adversarial-input, with reproducible, content-addressed evidence. Tools: run_litmus, verify_attestation.