Model Context Protocol MCP server
Commit Oct 27, 2026Provides a technical overview and implementation guidance for the Model Context Protocol specification to help engineers build and test MCP servers.
Also known as cyanheads/model-context-protocol-resources
- Stars
- 295
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- —
- Last commit
- Oct 27, 2026
- License
- Apache-2.0
What is Model Context Protocol MCP?
Model Context Protocol provides a technical overview and implementation guidance for the Model Context Protocol specification to help engineers build and test MCP servers.
Summary generated from the project's README
About Model Context Protocol MCP server
Engineers building and testing Model Context Protocol servers can use these resources to understand the technical specification and implementation requirements. The documentation provides a working model of the protocol, covering essential concepts like roles, wire formats, transports, and authorization. It specifically addresses the transition to a stateless protocol as defined in the 2026-07-28 revision, helping developers navigate changes from legacy versions.
Developers can utilize the @cyanheads/mcp-ts-core TypeScript framework to scaffold new projects using the bunx command. This framework includes specialized Agent Skills designed for coding agents to assist with the development lifecycle. These skills help automate tasks such as designing tool surfaces, auditing tool definitions, and performing security reviews for prompt injection or input sinks. The resources support various transport methods including stdio and Streamable HTTP to ensure compatibility across different client environments.
What you can do with it
- 1Design a new serverAsk an AI agent to plan the tool surface, resources, and service layer for a specific domain or API.
- 2Audit tool definitionsHave a coding agent analyze how tool descriptions and schemas will be interpreted by a language model.
- 3Perform security reviewsUse specialized skills to check a server for security gaps like prompt injection and authentication issues.
Written from the project's README. Every use case is checked against a quote from it.
Automated checks
Ran Nov 14, 2026Full report →Not verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachablecyanheads/model-context-protocol-resources is public and reachablegithub.com/cyanheads/model-context-protocol-resources ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)Apache-2.0spdx.org/licenses/Apache-2.0.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-10-27github.com/cyanheads/model-context-protocol-resources ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubThe MCP overview explains how MCP works: the roles, the wire format, what a server exposes, how servers ask for input, the transports, authorization, and what changed when the protocol went stateless in revision 2026-07-28. It is written for engineers who want a working model of the protocol before reading the specification.
Building a server? @cyanheads/mcp-ts-core is an agent-native TypeScript framework for building MCP servers on the current spec. Every scaffold ships Agent Skills that your coding agent follows to design, test, and release the server, and the server answers legacy and modern clients alike. Scaffold a project with:
bunx @cyanheads/mcp-ts-core init my-mcp-server
Looking for working servers? The servers I maintain are listed on my GitHub profile. The Fleet Map lists every hosted one with its tools, endpoint, and install snippets.
Guides
| Guide | Covers |
|---|---|
| MCP overview | Hosts, clients, and servers; messages and statelessness; tools, resources, and prompts; multi round-trip requests; notifications and caching; stdio and Streamable HTTP; authorization; extensions; security |
| Legacy vs. modern servers | Revision history from 2024-11-05, what changed in 2026-07-28, deprecations, era compatibility and detection, and a migration checklist for server authors |
Agent Skills for server builders
These skills ship with mcp-ts-core and are written for coding agents. They reference the framework's APIs, but the practices carry over to any SDK. Hand one to your agent and ask it to distill the guidance for your stack.
| Looking to… | Skill |
|---|---|
| Design an MCP server: tool surface, resources, and service layer for a domain or API | design-mcp-server |
| Build a good tool: naming, schemas, response design, errors | add-tool |
| Audit how your tool descriptions and schemas read to a model | tool-defs-analysis |
| Shape large or awkward responses: overflow handling, payload shaping, retrieval patterns | techniques |
| Test a server against live inputs, including adversarial ones | field-test |
| Review a server for security gaps: prompt injection, input sinks, auth, leakage | security-pass |
Official resources
| Resource | Link |
|---|---|
Specification (2026-07-28) |
modelcontextprotocol.io/specification/2026-07-28 |
| Changelog for this revision | Key changes |
| Deprecated features registry | Deprecated |
| Schema (source of truth) | schema.ts |
| Extensions | Extensions overview |
| Enhancement proposals | SEPs |
| Server registry | MCP Registry |
| Debugging | MCP Inspector |
Official SDKs: TypeScript · Python · Go · Java · Kotlin · C# · Rust · Swift · Ruby · PHP
Model Context Protocol MCP questions
What is included in the Agent Skills for server builders?
The framework provides specialized skills for designing servers, building tools, auditing tool definitions, shaping responses, testing, and security reviews.