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.
-
Deterministic, offline FMECA MCP server: event-sourced failure-mode ledger, fixed criticality matrix, mitigation discipline, readiness gate.
cargo Apache-2.0 -
Faults and Guards: failure modes a model commits doing measurement work, and the procedures against them, as data, prompts, tool schemas and an MCP server.
NOASSERTION -
Failures — deterministic MCP for AI-built software. 13 tools that make coding agents handle failure modes (idempotency, atomicity, timeouts, retries) via AST checks, not prompts.
pypi ★ 4 -
ejentum-mcpno update data
Exposes the four Ejentum cognitive harnesses (reasoning, code, anti-deception, memory) as MCP tools any agentic client can call. Drop-in scaffolding that catches LLM failure modes like sycophancy, hallucination, and reasoning shortcuts.
smithery -
discord-mcpgithub.com/jaimenbell/discord-mcpupdated this year
MCP server over the Discord REST API: 5 read tools always on, 7 write tools gated off by default behind an env flag. Typed errors for every failure mode.
pypi MIT ★ 1 -
ai-test-pilotgithub.com/Drzymek92/ai-test-pilotupdated this year
LLM-driven test generator: schema-validated scenarios → runnable pytest & Playwright tests, with typed-input construction, golden/characterization mode, failure triage, and an MCP server.
MIT ★ 1 -
ThumbGate Pre-Action Checks self-improve from ranked lessons and repeated failures, hard-block detected secret leaks, and block matches in strict mode.
npm MIT ★ 31 -
content-hooks-mcpgithub.com/closermethod/content-hooks-mcpupdated this year
Content Hooks MCP for creators and AI agents. Short-form mechanics: 10 hook styles with failure modes, script structures for 5 formats, retention engineering, CTA rules, hook + script auditing. Receipts over hype.
★ 1 -
A safety layer between a language model and a broker order. MCP server: seven tools over a swappable broker backend, with the failure modes tested rather than assumed.
MIT -
A hand-written MCP server at the protocol level plus a LangGraph agent over it, with four topologies run against the same task set and their failure modes measured.
MIT -
advanced-writergithub.com/angrysky56/advanced-writerupdated this year
A modular narrative engineering system that uses neurochemical pacing, Jungian depth psychology, structural paradigm selection, and automated pathology diagnostics to produce fiction that resists the default failure modes of AI-generated prose.
GPL-3.0 ★ 7 -
Denpex MCP server, diagnose GPU/ML training failures (CUDA OOM, NCCL timeouts, Xid faults, NaN loss) from Claude Desktop, Cursor, or Claude Code, with optional explicit zero-egress mode.
npm MIT -
agent-arena: evaluate an AI agent against deterministic failure-mode scenarios over REST, WebSocket, MCP or A2A, and get report.json + SARIF. The engine runs in-process; no model runs anywhere in the open core. Published as one self-contained bundle: no r
npm Apache-2.0 -
content-hooks-mcpupdated this year
Content Hooks MCP for creators and AI agents. Short-form content mechanics: 10 hook styles with failure modes, script structures for 5 formats, retention engineering, CTA rules, hook + script auditing. Receipts over hype.
npm MIT -
Make agent failures loud — deterministic checks that catch LLM agents claiming done when the work is placeholder, unbacked, stale, miscounted, or missing pieces. CLI + MCP server + watch mode + LangChain adapter + second-pass verification against the system of record, zero dependencies.
pypi MIT -
Benchmark the full voice-AI pipeline on your own audio — STT (batch + streaming), LLM TTFT, speech-to-speech, and TTS. WER by failure mode, keyterm recall, real-time-paced latency, cost columns, a CI regression gate, and an MCP server for coding agents.
MIT ★ 1 -
A governed GA4 semantic layer in dbt + MetricFlow on BigQuery, queryable by humans and by LLM agents over MCP. The README is a metric definition document: every metric states its definition, assumption, trade-off and failure mode.
MIT -
pdm-agent-mvpgithub.com/DISSIDIA-986/pdm-agent-mvpupdated this year
Multimodal PdM agent runtime: vibration (CWRU) + acoustic (MIMII-shaped) → LangGraph human-approval workflow → shared SQLite audit trail + MCP server + temperature-scaling calibrator. 110 tests, honest 0/30 ball-fault failure mode, second-modality F1=1.0 (synthetic internal-consistency smoke).
MIT -
agent-control-planegithub.com/sujithnsn953/agent-control-planeupdated this year
The systems layer around an agent, not another agent. Dual-era MCP server (2026-07-28 + legacy handshake), scoped tool grants with seven detectable failure modes, OTel GenAI tracing, governed memory, hybrid retrieval measured with recall@k, human-in-the-loop approval. Zero runtime deps.
MIT -
What's Allowedno update data
What your coding agent can do without asking you. Reads every settings file that contributes permission rules — managed policy, project, local, user — and reports the merged picture: which file each rule came from, which defaultMode wins, what proceeds unattended, and which rules the client accepts and then ignores (path rules on Write/Glob, unanchored allow globs, /path rules in user settings that anchor at the config directory, allow rules a deny rule reaches first). Four read-only tools: whats_allowed (full summary), permission_sources, rule_findings, unattended_surface. Read-only by construction — no child processes, no network, no writes; values of env entries are never read, only their names. MIT licensed. Free and MIT, with no telemetry and nothing held back. Built by Shift The Culture, who also publish paid kits for the failure modes this server surfaces: https://shifttheculture.media/agent-tools
smithery -
the-metis-fca-handbook-ai-harness-mcpno update data
# Metis FCA Handbook AI Harness — MCP Server An MCP (Model Context Protocol) server that integrates the Metis FCA Handbook AI Harness into Claude Desktop and other MCP-compatible agents. **Tool name:** `evaluate_fca_handbook_applicability` Use this tool to analyze FCA Handbook applicability for a product, service, or firm. Returns structured compliance reasoning with verbatim handbook citations. ## Why choose The Metis FCA Handbook AI Harness? - **Efficiency** Multiplies effectiveness of compliance advice. Saves £80+ in token fees per Harness run. - **Deals with "Hard Problem"**, in which LLMs' token incentives prioritise training data so results are unreliable. - **Verbatim citations** Quotes verbatim entries from the FCA Handbook. Other AI systems struggle to do this. - **Matches real-world need** You do not have to start off knowing which sections you are looking for. Carries out structured searches across all 10,000+ FCA Handbook entries. - **Secure Design** Harness compartmentalisation, one-shot structure, and statelessness fits natively with emerging AI agent security standards such as OWASP Top 10 for Agentic Applications 2026. - **AI Accessible** Integrates easily with AI workflows and agents. Your LLM can use it as a tool. ## Design Principles The Harness is built on proven principles: - **Stateless** — Each request is independent; no session coupling - **One-shot** — Complete analysis in a single call; no multi-turn state - **Clear contract** — Explicit input/output schemas for easy integration - **Hosted** — Single canonical source; no version drift or stale data ## Use Cases ### Compliance AI Platforms Embed FCA reasoning as a service within your compliance platform. Users ask natural-language questions; your platform calls the Harness and presents structured reasoning. ### AI Agent Workflows Agents building compliance workflows can include FCA Handbook reasoning as a composed tool—no external API calls, just MCP configuration. ### Compliance Review Automation Integrate into document review or due-diligence pipelines. Automatically screen new rules against FCA applicability. ## For Regulatory Screening FS firms screening MCP servers will find: - **Simple model:** Stateless, no hidden state, no background jobs - **Clear contract:** Explicit input schema, output schema, error modes - **Transparent updates:** Version pinning; no automatic upgrades - **Single point of failure:** If the Harness is down, it is obvious; no cascading config issues - **Audit-friendly:** All calls logged centrally, not distributed
smithery -
agentedgithub.com/frane/agentedno update data
A text editor for LLMs, not humans. Take ed, the line editor that nobody has voluntarily used since about 1975, and rebuild it for an environment where the typing user is a language model. Short verbs, line addresses, no modes, no TUI. What an editor optimises for changes when the user is the model: round trips per task, tokens per command, an editing buffer with a long memory, and an undo tree that remembers the branches the agent abandoned, because that's often where the interesting work was. ## Features - **Fewer round trips.** Read-before-Edit is unnecessary; on conflict the response carries the new content so you reconcile in one call instead of pre-reading every time. - **Branching undo.** Walked-back work stays addressable instead of being thrown away when you pick a different path. - **Three-way merge.** Concurrent agents get a structured conflict response instead of a silent overwrite. - **Atomic batches.** Multi-file refactors run all-or-nothing instead of leaving half-applied state on failure. - **Cross-file moves and regex replace as primitives.** Operations the built-in tools can't express cleanly become single calls. - **Drift detection.** External edits to an open file are folded into the tree instead of being clobbered by the next write. - **Inline diagnostics.** Type errors and lint findings surface on save, not at the next build many edits later. - **Cross-session memory.** Per-file notes persist between sessions and surface inline on the next open. - **Audit log.** Every operation recorded with actor and timestamp, so two agents in one workspace can't argue about who moved the head. ## Multi-workspace (v0.4.0) One MCP server registration handles every project you mention. Tool calls route to the workspace owning the absolute path you pass; each project gets its own SQLite, its own undo tree, its own LSP daemon — lazily spawned on first touch. Desktop hosts (Claude Desktop, Codex Desktop, Cursor) install the server once globally and use it across projects without per-project config. ## Install `ae` must be installed locally; this MCP bundle registers `ae serve` as a stdio server. ```sh brew tap frane/tap && brew install agented # or curl -sSL https://raw.githubusercontent.com/frane/agented/master/install.sh | sh ``` Pure Go, single static binary, Apache 2.0. Source: https://github.com/frane/agented
smithery -
fmeca-mcpgithub.com/matt-cochran/fmecaupdated this year
Stdio MCP server wrapping the fmeca kernel: a deterministic, offline structured-FMECA engine exposed as a command/query MCP tool surface.
cargo Apache-2.0 -
Open-source docs you zoom through: a C4-style model in YAML, step-by-step flows, visualization for performance stats and FMEA creation
GPL-3.0