# Koma Core MCP [Health: Active] **Category:** πŸ’° Finance & Fintech **Repository:** https://github.com/swnotmetal/Project-Koma **GitHub Stars:** 12 **Views:** 0 **Installs:** 0 **Upvotes:** 0 **Directory Page:** https://allmcps.com/mcp/koma-core-mcp ## Description Protected RAG storage β€” agents discover public metadata, content only by access token. ## Claude Desktop Quick Installation Install path detected from listing signals. Uses `npx` (confidence: high): ```json "mcpServers": { "koma-core-mcp": { "command": "npx", "args": ["-y","koma-miko"] } } ``` ## Documentation & README # Koma ### Miko: required Skill checks for Claude Code and Codex Keep reminding your coding agent to read a required Skill before editing? **Miko** checks for observable Skill reads before protected edits. When evidence is missing, it pauses the action and tells the agent what to load before retrying. It runs locally, makes no LLM calls, and is free and open source. The current focus is **Claude Code and Codex CLI**. Miko is part of Koma; the separate **Gate, Scout, and Core** packages cover AI application input, request limits, and retrieval. You do not need them to use Miko. ```bash npm install -D koma-miko@alpha npx koma-miko init --host claude ``` Edit the generated `miko.json` to name your project's Skills and protected paths, then start a new Claude Code session. **Using Codex?** Follow the [Codex setup and one-time Hook review](https://github.com/swnotmetal/Project-Koma/blob/HEAD/packages/koma-miko/README.md#codex-setup). Codex CLI is a Technical Preview; Desktop requires prior CLI activation.

Koma

License CI koma-miko koma-gate koma-scout koma-core
Miko live demo Gate benchmark koma-gate downloads MCP server

δΈ­ζ–‡η‰ˆ

β–Ά Try Miko's guided terminal replay β€” plus Gate, Scout & Core, no signup.

--- ### Miko for Claude Code and Codex

Koma Miko

Coding agents can say they loaded a required Skill or ran a test. Miko does not treat that claim as evidence. At supported local host Hooks, it compares observed Skill loads, reference reads, tool actions, and completion checks with a project-owned `miko.json`. If an agent tries to edit before satisfying the spec, Miko can return a denial and a short recovery instruction. It cannot inspect hidden model context, prove that a model understood a Skill, or verify events the host never exposes. In Claude CLI guided mode, a genuine scope exception opens one visible Miko choice: allow that exact retry once, or keep the current boundary. ```bash npx --yes koma-miko@alpha demo # deterministic; no API key npx --yes koma-miko@alpha probe --host claude # isolated adapter check; no model npx koma-miko init --host claude # after local installation ``` > **Codex CLI is a Technical Preview; Codex Desktop requires prior CLI > activation.** The CLI requires a one-time `/hooks` review before project > Hooks run. Miko promotes `enforce` on Codex because the current Hook API > cannot open a native `REVIEW` choice; review therefore degrades to a > recoverable pause/deny. Desktop-only onboarding is not a supported alpha > path. An offline `probe` proves adapter logic, not live Hook activation. [Miko README β†’](https://github.com/swnotmetal/Project-Koma/blob/HEAD/packages/koma-miko/README.md) Β· [Current host support β†’](https://github.com/swnotmetal/Project-Koma/blob/HEAD/packages/koma-miko/README.md#host-support) Β· [10-second web replay β†’](https://koma-demo.swbuilds.workers.dev) [Miko replay GIF](https://github.com/swnotmetal/Project-Koma/blob/HEAD/docs/assets/miko-terminal-demo.gif) Β· [Still image](https://github.com/swnotmetal/Project-Koma/blob/HEAD/docs/assets/miko-terminal-demo-poster.png) β€” browser simulation, not a live host recording. Claude Code is the primary alpha workflow. Codex CLI has a verified narrow recovery flow with the activation limits above. Gemini is outside active development; Copilot adapter work is paused pending a real tester. Existing adapters and their dated results remain available in the [adapter documentation](https://github.com/swnotmetal/Project-Koma/blob/HEAD/packages/koma-miko/README.md#other-adapters). **Why use Miko instead of writing a Hook?** A small native Hook is enough for a single fixed check. Miko packages project Specs, observed-read tracking, post-compaction reload requirements, recovery messages, and completion evidence so you can maintain those checks together. Neither approach proves the model understood the instructions. See [when Miko helps](https://github.com/swnotmetal/Project-Koma/blob/HEAD/packages/koma-miko/README.md#when-to-use-miko). --- ### Four Boundaries | Boundary | Failure mode | What Koma checks | Package | |---|---|---|---| | Coding agent | Required Skill or completion check skipped | Host-observed preparation, action scope, and evidence | `koma-miko@alpha` | | User β†’ LLM | Prompt injection / jailbreak | Semantic scope before the application model | `koma-gate` | | Request perimeter | Audio abuse / flooding | Validation, rate limits, and geo rules | `koma-scout` | | Retrieval | Data enumeration / scraping | Split index from content; token-gate retrieval | `koma-core` | Different attacks cross different boundaries. Koma provides a small primitive for each one. --- ### What Koma Is β€” and Isn't **Is**: small composable packages Β· usable independently Β· explicit failure modes Β· deterministic checks where the host exposes evidence **Isn't**: a model Β· an agent framework Β· proof that a model understood its instructions Β· a replacement for authorization Β· a complete security boundary by itself --- ### Benchmarks #### Miko alpha evaluation Miko is deterministic, so its useful numbers are verifier cost and end-to-end Hook behaviorβ€”not a generic score for model intelligence. | Signal | Observed result | |---|---| | Offline host conformance | Claude, Codex, Gemini, and VS Code Copilot each reproduce `DENY β†’ observed Skill β†’ ALLOW`; ledger fixtures reject prompt/code/tool-response persistence | | Local verifier scale | 1,000 Agent Specs: **1.34 ms p95** per action; 10,001 indexed evidence events: **0.0041 ms p95**; restore 1,000 evidence events: **1.52 ms p95** | | Claude Code live + hand-test | One 100-Skill / ~20k-context run passed. A natural Chinese request drove a three-file rewrite through 3 Agent Specs and 28 observed events. Separate Haiku 4.5 runs verified visible `Allow once` and `Keep current scope` policy-exception paths | | Codex CLI Technical Preview | Fixed live recovery completed `DENY β†’ Skill/reference β†’ edit β†’ COMPLETE`; an interactive CLI 0.152.0 hand-test visibly rendered Miko active, recovered, and COMPLETE. Desktop also enforced after CLI activation, but onboarding and completion visibility remain unsuitable as a primary path | | DeepSeek Harness smoke | **3/3** narrow packed-artifact recovery runs passed; 19.425 s mean model phase | The scale row is a 2026-08-27 reference run on Node 24.19 / Windows; rerun it with `npm run eval:scale -w koma-miko`. Context tokens never enter the verifier. The paid samples are deliberately small and **do not establish general model, long-context, or editor reliability**. See the [scale record](https://github.com/swnotmetal/Project-Koma/blob/HEAD/docs/evals/miko-scale-alpha.md), [Claude record](https://github.com/swnotmetal/Project-Koma/blob/HEAD/docs/evals/miko-claude-haiku-alpha.md), [host-adapter record](https://github.com/swnotmetal/Project-Koma/blob/HEAD/docs/evals/miko-host-adapters-alpha.md), and [DSH record](https://github.com/swnotmetal/Project-Koma/blob/HEAD/docs/evals/miko-dsh-alpha.md). #### Koma Gate live-model benchmark I threw **1,769 real prompt-injection attacks** at Koma Gate in fail-closed mode, using real providers β€” not mock adapters. | Provider | Recall | Precision | False Positives | |----------|:------:|:---------:|:---:| | DeepSeek (deepseek-chat) | **98.8%** | **100%** | **0** | | Google (gemini-2.5-flash) | 96.2% | **100%** | **0** | **Chinese attack set**: 100% recall Β· 100% precision Β· 0% FPR across 8 categories. > **Can you break it?** [Open an issue](https://github.com/swnotmetal/Project-Koma/issues) with an attack Koma misses. β†’ [Full methodology](https://github.com/swnotmetal/Project-Koma/blob/HEAD/BENCHMARKS.md) --- ### Application-side quick start: Gate ```ts import { createGeneralKnowledgeGuard } from 'koma-gate'; const guard = createGeneralKnowledgeGuard({ llm: { apiKey: process.env.GEMINI_API_KEY }, }); app.post('/api/chat', guard.middleware(), async (req, res) => { // Only in-scope requests reach your model res.json({ reply: await chat(req.body.message) }); }); ``` ```bash git clone https://github.com/swnotmetal/Project-Koma cd Project-Koma && node demo/server.js curl http://localhost:8080/self-test ``` --- ### Application-Side Packages **`koma-gate`** β€” Prompt injection firewall. LLM-based scope classifier that blocks jailbreaks, off-topic requests, and instruction overrides. Supports OpenAI, Anthropic, Google, DeepSeek, and local Ollama models. [README β†’](https://github.com/swnotmetal/Project-Koma/blob/HEAD/packages/koma-gate/README.md) [Try Gate in the browser demo β†’](https://koma-demo.swbuilds.workers.dev) Koma Gate blocking a prompt injection in real time **`koma-scout`** β€” Perimeter protection. Rate limiting, audio upload validation, geo allowlisting. Cheap checks before expensive AI work. [README β†’](https://github.com/swnotmetal/Project-Koma/blob/HEAD/packages/koma-scout/README.md) Koma Scout perimeter checks **`koma-core`** β€” Protected RAG storage. Public search index, private content, opaque HKDF-derived tokens. *Discovery is not authorization.* [README β†’](https://github.com/swnotmetal/Project-Koma/blob/HEAD/packages/koma-core/README.md) Koma Core split-store Each package works standalone. A typical application checks cheap request limits with Scout before Gate classification, then uses Core where protected retrieval is needed. See the [architecture map](https://github.com/swnotmetal/Project-Koma/blob/HEAD/ARCHITECTURE_MAP.md). **Application-side MCP servers** β€” these belong to Gate and Core, not Miko: - `koma-gate-mcp` β€” `classify_input` tool for prompt-injection checks. [README β†’](https://github.com/swnotmetal/Project-Koma/blob/HEAD/packages/koma-gate-mcp/README.md) - `koma-core-mcp` β€” `search_docs` + `retrieve_doc` for protected RAG retrieval. [README β†’](https://github.com/swnotmetal/Project-Koma/blob/HEAD/packages/koma-core-mcp/README.md) ```json { "mcpServers": { "koma-gate": { "command": "npx", "args": ["-y", "koma-gate-mcp"] }, "koma-core": { "command": "npx", "args": ["-y", "koma-core-mcp"] } } } ``` --- ### Using an agent to set up Miko? Tell it: > "Read the Miko README, then help me configure koma-miko for my existing Skills and protected paths. Use the setup for my host and explain any activation step I must complete." For a coding-agent repository, install Miko and run `npx koma-miko init --host claude` or follow the [Codex setup](https://github.com/swnotmetal/Project-Koma/blob/HEAD/packages/koma-miko/README.md#codex-setup). Edit the generated `miko.json` to name the Skills, paths, and completion evidence that matter to the project. Miko does not install the Skills themselves. See [llms.txt](https://github.com/swnotmetal/Project-Koma/blob/HEAD/llms.txt) for documentation entry points. Miko uses host Hooks; it does not require an MCP server. --- ### Trust & Safety - **Minimal dependency surface.** Miko, Gate, and Core have no third-party runtime dependencies; Scout declares Express as a peer. - **No model-output execution.** Miko observes host events; Gate, Scout, and Core classify, rate-limit, or store. None executes generated code. - **Package-specific failure behavior.** Gate defaults to fail-open and can use `failOpen: false`. Miko follows each Agent Spec's mode; an enforce-mode missing-evidence check denies the applicable action. - **CodeQL on every push.** Targets OWASP LLM01. - **MIT licensed.** β†’ [Security policy](https://github.com/swnotmetal/Project-Koma/blob/HEAD/SECURITY.md) Β· [Known limitations](https://github.com/swnotmetal/Project-Koma/blob/HEAD/SECURITY-HARDENING.md) Β· [Contributing](https://github.com/swnotmetal/Project-Koma/blob/HEAD/CONTRIBUTING.md) --- Koma comes from Komainu ("η‹›ηŠ¬"), the stone guardian lions of Japanese Shinto shrines. Three deployed defense layers, each standalone, plus the Miko alpha agent-contract boundary. Patterns distilled from production, not papers. [License](https://github.com/swnotmetal/Project-Koma/blob/HEAD/LICENSE)