Skip to content
Agentmart
Submit a tool

mcp-application

MCP as a product protocol for generated applications: tools, resources, schema validation, and structured errors.

Skill by Literate AI maintainers <literate-ai-maintainers@users.noreply.github.com>

IndexedApache-2.0★ 28
Write a review

Chalk Labs

Need help deploying mcp-application?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/jordanhubbard-mcp-application-da1c4f

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

MCP as a product protocol for generated applications: tools, resources, schema validation, and structured errors. Use for Literate AI workflow tasks that expose MCP in a page, process, or service. Do not use for operator catalog setup.

Use cases: productivity, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/mcp-application

What's inside

Skill name
mcp-application
Description
MCP as a product protocol for generated applications: tools, resources, schema validation, and structured errors. Use for Literate AI workflow tasks that expose MCP in a page, process, or service. Do not use for operator catalog setup.
License
Apache-2.0
Metadata
author: Literate AI maintainers <literate-ai-maintainers@users.noreply.github.com>
Instructions
61 lines in SKILL.md · view on GitHub

Files (2)

  • SKILL.md
  • SKILL.md · webmcp/

SKILL.md (first lines)

# MCP application protocol

MCP is one protocol. Nested skills are deltas of this parent: they add a leaf
(transport, in-page actuation, or language ecosystem) without restating hygiene,
identity pins, or secret rejection.

## MCP is the first adapter of the unified IPC surface

MCP is one **adapter of the framework's IPC-surface notion** (ADR 0029), not a
standalone protocol — its peers `rest-application` (OpenAPI/JSON-Schema) and
`grpc-application` (`.proto` + reflection) are the other two adapters of the same
contract. Every adapter exposes the same four surface properties, which this
parent already requires for MCP and only names generically here: a **declared
schema** (schema-validated input, typed output, structured errors — below); a
**served self-description** a consumer or verifier can fetch (for MCP, the
tool/resource listing; for REST, `/openapi.json`; for gRPC, server reflection);
a **conformance acceptance** that fails a surface whose served description does
not match its declared schema (the verifier-owned
`literate-ai/ipc-surface-conformance-acceptance@1` oracle, protocol tag `mcp`,
launched via `--litai-serve`); and a **semantic version + compatibility promise**
reusing the framework's existing `CompatibilityPromise`. An MCP surface is not a
new entrypoint kind — a deployment unit exposes it. Do not invent a second schema,
self-description, or version mechanism when a surface already declares one.

## One protocol, four distinct leaves

Keep these leaves distinct. Do not collapse them in generated source or in this
recipe:

1. **Personal operator catalog** — MCP servers a *user* opts into for the coding
   session. That catalog is not product policy and must not appear as a generation
   requirement.
2. **Project MCP hygiene** — servers the *project* versions under `mcps/<id>/mcp.md`
   with identity pins, no secrets, and no duplicate ids. Catalog hygiene is not an
   application entrypoint.
3. **Embedded application MCP** — a generated process exposes tools and resources as
   product behavior (stdio or HTTP/SSE). Nested backend skills own the process and
   store; they inherit this protocol chapter.
4. **WebMCP** — the same protocol *in the page*: JavaScript functions or HTML tools
   registered for in-browser agents. Nested WebMCP skills inherit this chapter and
   add in-page actuation. WebMCP is not a second protocol and is not a backend
   integration that bypasses the page.

## Tools, resources, and errors

Expose tools (actions and parametrized queries the caller invokes) separately from
resources (readable context addressed by URI). Validate every tool's input against an
explicit schema before executing it, and reject invalid input with a structured error
rather than executing a partially-valid call. Every tool has a concise description
covering what it returns and its parameters' constraints. Errors returned to an MCP
caller are structured, not a bare string. Do not invent a second tool schema language
when the specification already names MCP.

## Hygiene this parent owns

Pin MCP identity the same way skills pin identity: a stable id, a semantic version,
and exact bytes after modify. Reject secrets in tool descriptions and sibling
generated metadata. Do not reuse reserved operator ids (`jira`, `slack`, `outlook`,
`registry`) as product tool or server names. Nested skills inherit these rules; they
must not copy them.

From jordanhubbard/literate-ai, Apache-2.0.

More from jordanhubbard/literate-ai

Signals

Stars
28
Forks
10
Open issues
26
Last commit (this folder)
Sep 30, 2026 (1 month ago)
Last release
Oct 12, 2026 (20 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
Apache-2.0
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

Same skill elsewhere

Copies of this SKILL.md published by others; search shows only one of them.

More from Literate AI maintainers <literate-ai-maintainers@users.noreply.github.com>

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get mcp-application running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying mcp-application?

CallGet help
Cite this page

mcp-application skill. Agent Mart. https://agentmart.dev/skills/jordanhubbard-mcp-application-da1c4f (last updated 2026-10-24)