MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
wispgithub.com/erhnysr/wispactive
Independent signal and participation tool for the Technocore network: paste a did:key to see what the network's own data says, every signal with a proves/doesn't-prove note.
MIT -
Attestix - AI agent identity and signed evidence tooling. DID-based identity, W3C Verifiable Credentials, governance evidence, delegation chains, and reputation scoring.
pypi Apache Software License -
MCP server for Kynver — verify AI agents, check trust scores, resolve DIDs, register agents, and complete ownership challenges or KYC directly from Cursor, Claude Desktop, or VS Code.
npm MIT -
tcxc-mcp-servergithub.com/TelecomsXChangeAPi/tcxc-mcp-serverupdated this year
MCP server for the TelecomsXChange (TCXC) API — give Claude, OpenAI, Cursor, and any MCP client AI access to wholesale voice termination, SMS, DID numbers, eSIM, and HLR lookups. 55 telecom tools.
MIT ★ 3 -
Runtime observability for AI agents — see exactly what your agent did, why it did it, and what it cost.
pypi MIT -
MCP server (2025-11-25, Streamable HTTP): tamper-evident log of what an AI agent did. "Alexa, what did my agent do yesterday, and was the log changed?" Voice simulator + MCP App card.
MIT -
An MCP server that gives a coding agent a record it did not write: what actually changed on disk and in git, held against what the agent said it did.
MIT -
Agent-native Stata bridge for empirical research — run DiD/IV/RDD and publication-ready tables from Claude Code, Jupyter, or VS Code on one token-economy result schema, with StatsPAI cross-validation. | 面向实证研究的智能体 Stata 桥接器——在 Claude Code、Jupyter、VS Code 中用一套省 token 的结果格式跑 DiD/IV/RDD 与出版级表格,并支持 StatsPAI 跨栈交叉验证。
MIT ★ 55 -
AI-first CLI and MCP server for Zabbix 6.4+: current problems, host investigation, why an alert did or did not arrive, metrics and maintenance windows, for Claude Code and any MCP client. Bounded JSON output. One setting, allow_write (on by default), decides whether the agent applies changes or a person approves a plan. Single Go binary.
Apache-2.0 -
VERAX body: an MCP server that decides, records and proves what an agent did.
npm Apache-2.0 -
We check whether your agent has an alibi for what it did.
MIT -
Auragithub.com/ojuschugh1/Auraupdated this year
Your AI remembers what it did, proves it, and gets smarter every session
MIT ★ 11 -
macrofactor-mcpgithub.com/NasserAlbusaidi/macrofactor-mcpupdated this year
Am I recovered? Did my fueling work? MCP server combining MacroFactor nutrition with Garmin training data — a local fueling & recovery decision engine.
MIT ★ 2 -
MCP connector for Fabric Verify: signed, publicly logged records of what your AI agent did, with proofs anyone can check offline.
Apache-2.0 -
When did Codex / Claude Code / Grok last reset usage limits? Terminal table + MCP server, backed by whenaireset.com
npm MIT -
code-tribunalgithub.com/arun3676/code-tribunalupdated this year
Did the AI build what you asked? Multi-agent intent-conformance review for AI-generated code — CLI, MCP server, live courtroom demo
MIT -
MCP server that appends what you did to the checklist of the newest log file in a directory
MIT -
MCP server for Vocabit — let an AI assistant build flashcard sets inside a real language-learning app and read back how the learner actually did.
npm MIT -
Understand what your AI coding agent actually did. A clarify-before / verify-after layer for Codex, Claude Code and OpenCode.
npm MIT ★ 1 -
Local, private "where did I write that?" search for your notes, built for coding agents: BM25 + local Ollama embeddings, MCP server, disk-verified results.
MIT ★ 1 -
agent-receipts-mcpgithub.com/dhanushs1912-svg/agent-receipts-mcpupdated this year
Prove what your AI agents did - EU AI Act-ready audit trails: signed, tamper-evident receipts with offline verification. MCP server + HTTP API.
MIT ★ 1 -
Your EXTREMELY agency of AI agents in any AI app (Claude Desktop, Cursor, VS Code, Claude Code): what agents did, approvals, requests, jobs, reports. Needs the Extremely Autonomous Agents app.
npm AGPL-3.0-only -
Audit trail and traceability for AI agents. See what every coding agent did and check each tool call against policy.
npm MIT -
See what your team did with AI: each session, tool call and dollar traced to a person. Self-hosted, Apache 2.0, REST + MCP.
Apache-2.0 ★ 1 -
MCP server: ask Claude "why did leads drop?" across GA4, Search Console and HubSpot. 7 diagnostic tools, demo mode, read-only. Built by zov911.com
-
Your agents, on a team with you: one MCP server joins any MCP-capable agent to a musterd team. Every piece of work shows who did it and who accepted it.
npm MIT -
MCP server for Vocabit: an AI assistant builds a flashcard set inside a real language-learning app, then reads back how the learner actually did. Ships with a no-backend demo mode.
MIT -
What a checkout you did not write tells your agent to do — instruction files, hook commands, declared MCP servers, and repo-shipped skills. Read-only MCP server.
MIT ★ 1 -
Undo and verify what your AI coding agent did: rollback, test-proof gate and token trimming for Claude Code, Codex and OpenCode (CLI + MCP server)
Apache-2.0 ★ 4 -
An MCP server that answers "did my deploy actually work?" — catches pages that return 200 and render nothing, broken links, stray noindex, and staging that never updated. Zero config, no API key.
MIT -
signalgithub.com/LSUDOKO/signalupdated this year
Signal is a Slack agent that makes work chat less overwhelming and more understandable for people with ADHD and autism—using AI to translate social tone, smart search to answer "what did I miss?", and calendar integration to protect focus time.
MIT -
Which laws say your AI agent must tell people it is an AI, and did this conversation do it? 47 rules (US, EU, UK, Canada, Australia), each quoting the official text. CLI, library, MCP server.
npm MIT -
The system of record for what an AI agent believed and did. Append-only beliefs with evidence, contradictions kept, risky actions gated behind a named human. MIT, self-hosted.
MIT -
Free MCP server that lets Claude work with FL Studio: read and edit projects, compose, mix, render and explain what it did - an ease-of-use tool for every music lover. Windows. Beta.
GPL-3.0 ★ 1 -
Agent-first open-source ERP for Swiss SMEs: 852 MCP tools, one local file. Sales, purchasing, payments, stock, projects, books and MWST; your agent runs it, the Studio shows what it did. MIT licensed.
npm MIT -
A local-first MCP server that answers, from real data on your machine: How many tokens have I used, from which client, model and session — and what did it cost?
MIT ★ 1 -
The deterministic trust layer for autonomous AI coding agents. Compiler-grounded verification of what an agent actually did vs. what it claimed, blast radius mapping, and persistent contract memory.
npm MIT -
A Model Context Protocol server that exposes a self-hosted Bamboo instance to AI assistants, Claude, Cursor, Windsurf, or any other MCP client. Ask "who deployed CHECKOUT to staging and did it pass?" instead of clicking through the Bamboo UI. Read-only by design. mcprush.com-4e3bcff355374530c531fe0a
Apache-2.0 ★ 1 -
whats-inherited-mcpgithub.com/stcmain/whats-inherited-mcpupdated this year
MCP server that shows what a checkout you did not write tells your agent to do — instruction files, hook commands wired to agent events, MCP servers it declares, and the skills and subagents it ships.
npm MIT -
Companies screen you with AI. Screen them back. Let your AI assistant find jobs straight from employers' own ATS: is the job still open, is the company real, did the posting you applied to close. Résumé stays on your machine. MCP server, MIT.
pypi MIT ★ 15 -
maestro-career-studiono update data
Gives your AI assistant your career history: it saves and scores job postings, tailors a resume to each one from things you actually did, renders the PDF, and tracks your applications. Needs Maestro Career Studio running on your computer: uv tool install --python 3.12 maestro-career-studio, then maestro start. 88 tools; your data stays on your machine.
smithery -
Verify what an AI agent did on someone's behalf. Resolves a signed agent receipt at a neutral registry, showing what was asked, what was done, what money moved, and what commercial influence the operator disclosed. Read only, no API key.
npm MIT -
Did your scheduled GitHub workflows run when they should have? Late, missed, failed and green-but-empty runs, workflows GitHub disabled, and the 60-day countdown, from your repo's own run history with your own token. CLI, GitHub Action, MCP tool.
MIT -
Which laws say your AI agent must tell people it is an AI, and did this conversation do it? 47 rules across the US, EU, UK, Canada and Australia, each quoting the official text. Checker as CLI, library and MCP server. Zero dependencies.
npm MIT -
Find out what your app did, from its logs to its threads, and ask AI about it (GitHub Copilot or Claude Code, MCP): multi-GB logs, live tail, filters, request tracing, exceptions, thread dumps, GC logs, heap and crash dumps, JVM profiling, SSH. Windows, macOS, Linux.
MIT ★ 2 -
Keep code understandable. The evidence layer between the agents that write code and the people who stand behind it: where to look first, who can look, what a change did to the structure and what nobody maintains. Offline, zero dependencies, with the evidence for every claim.
MIT -
CogmemAi is the safe, persistent memory that keeps Ai in line. It remembers what matters across every session, stops any action that breaks the rules you set before it happens, and reviews what your Ai did against what you asked for. Rule packs: npx cogme
npm MIT -
Scores a Claude Code harness from the transcripts already on your machine: 6 stats of what your agent actually did, computed locally with no LLM calls. Then it runs a reproducibility gate over its own numbers and prints which ones failed and which it could not test at all.
MIT ★ 1 -
See what your coding agent actually did, and make it fix the tests it broke: a floating pal for Claude Code, Codex and any agent that checks every file changed, command run and test result, plus an MCP server any AI tool can ask for the evidence. New: tal
npm MIT -
hob_hunt_mcpgithub.com/Corykidios/hob_hunt_mcpupdated this year
A g-search and a fetcher had a little baby, yes they did; there where three in the family, because that's the magic number. More like free is the magic number, am I right?! Enjoy a low effort appropriation of two well crafted servers into one poorly named and even more poorly understood server.
★ 3 -
Your agent wants to touch prod. Shadow Run lets it break a clone first: every risky plan runs on a full clone of prod in a microVM on Vultr, and a person approves what it really did, not what it says, before it reaches the real database.
MIT -
Kanbanto is a task board you share with your AI. It reads the board, so it knows what needs you today without being told. It works there too: it picks up cards, does the work and writes back what it did. You decide what's next, and what counts as done. Self-hostable, fair-code.
NOASSERTION ★ 3 -
A Model Context Protocol server that drives Aseprite headlessly and measures what it drew: 156 tools to draw, shade and animate pixel art, and results that say what each call did. An agent that cannot see the screen can diff two frames, score palette conformance, and verify a loop. Exports PNG/GIF/sprite sheets, Godot and Minecraft packs.
MIT ★ 40 -
Recallgithub.com/adityapall1/Recallupdated this year
Recall is a personal knowledge base you can talk to. Save notes, snippets, and links, then search them by meaning rather than exact wording. Recall runs as an MCP server, so it works directly with your AI assistant, so you can ask Claude "what did I save about hash tables?" and get an answer from your own notes. There is also a CLI and web app.
MIT ★ 1 -
agent-robustnessno update data
Run an AI agent against 5 adversarial attacks (credential leak, metric fabrication, false halt, runaway spend, log falsification), scored on what the harness observed it do '—' never its self-report. Free local self-diagnostic: runs on your own machine, in a temporary working directory, not a security sandbox. Optional paid tier runs the same five server-side in fresh single-use E2B microVMs (isolation enforced by E2B) and returns an ed25519-signed badge. Each attack is PASS, FAIL, or INDETERMINATE; an attack the agent did not demonstrably encounter is never a PASS, and a run containing an INDETERMINATE is never signed. A badge attests one artifact's observed behavior on one date. Not a safety certification.
smithery -
Vorim AIno update data
Identity, permissions, and signed audit trails for AI agents. Every agent gets an Ed25519 cryptographic identity, scoped permissions, and an append-only audit trail. Permission checks run in under 5 milliseconds. Audit events export with a SHA-256 manifest for tamper-evident, offline verification. Trust scores and identity verification are publicly accessible via REST endpoint. Anchored to the IETF Internet-Draft for VAIP (the Vorim Agent Identity Protocol). **17 tools** covering agent registration, ephemeral agents (W3C did:key), permissions, audit events, audit export, credential delegation, scoped token issuance, and public trust verification. **Get an API key:** [vorim.ai/settings/api-keys](https://vorim.ai/settings/api-keys) **Docs:** [vorim.ai/docs](https://vorim.ai/docs) **Open protocol:** [github.com/Kzino/vorim-protocol](https://github.com/Kzino/vorim-protocol)
smithery -
Contextualno update data
## Contextual — Persistent Memory for AI Coding Agents AI coding agents forget everything the moment your session ends. **Contextual** fixes that by giving any MCP-capable AI client — for developers who use multiple AI coding tools at once — a real, persistent memory of your codebase: not just what the code says, but what changed and why. It's a **local-first MCP server**: your code, your index, and every query stay on your machine. The only network call is a lightweight, infrequent license check — no code, ever, leaves your device. Under the hood, Contextual builds a real understanding of your repository, not a text-similarity guess: - **Structural search** — tree-sitter AST parsing + call-graph resolution, so agents reason about how your code actually connects. - **Bitemporal history** — tracks not just when your code changed, but when the AI learned about it. Ask "did this bug fix happen before or after the agent's understanding of this function?" and get a real answer. - **Git-aware retrieval** — temporal, history-grounded context instead of a stale snapshot. - **Fast, self-contained** — index a 50k-line repository in about 5 minutes, then query it in ~94ms median, from a single ~0.8GB install. This extension connects Claude Desktop (or any MCP-capable client) to a Contextual installation already set up and activated on your machine — it does not provision Contextual from scratch.
smithery -
cogmemgithub.com/writerslogic/cogmemupdated this year
A self-improving, verifiable memory layer for AI coding agents.
pypi Apache-2.0 -
trust-mcpgithub.com/synpareia/trust-mcpupdated this year
Identity and trust tools for AI agents — MCP server
Apache-2.0 -
Free, open reputation registry for AI microservices. Agents check trust scores before calling any service.
npm Apache-2.0