Borrowing it
Nothing to install: this file belongs to Roee-Tsur/mcp-spec-check. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/Roee-Tsur/mcp-spec-check/main/CLAUDE.mdgit clone --depth 1 https://github.com/Roee-Tsur/mcp-spec-checkWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/roee-tsur/mcp-spec-check/claude-md)<a href="https://agentmods.dev/instructions/roee-tsur/mcp-spec-check/claude-md"><img src="https://agentmods.dev/badge/instructions/roee-tsur/mcp-spec-check/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/roee-tsur/mcp-spec-check/claude-md"><img src="https://agentmods.dev/badge/instructions/roee-tsur/mcp-spec-check/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.03441 | $0.03441 |
| Opus 5 | $0.01721 | $0.01721 |
| Sonnet 5 | $0.00688 | $0.00688 |
| Haiku 4.5 | $0.00344 | $0.00344 |
Grade A, and why
mcp-spec-check CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
mcp-spec-check
Zero-install CLI (npx mcp-spec-check <url>) that black-box-probes a remote MCP server and reports whether it's ready for the MCP spec release of 2026-07-28. Second deliverable: a scan of the public registry ("X% of public MCP servers aren't ready") published as a writeup.
Commands
npm run dev <url>— run the CLI from source (tsx)npm run build— compile to dist/npm test— vitestnpm run typecheck— tsc --noEmit
Requires Node ≥20 (uses global fetch); develop and install under Node 22.
Architecture
src/index.ts— CLI entry: arg parsing, check runner, exit codes (0 ready / 1 fail / 2 error)src/client.ts— JSON-RPC-over-HTTP probe helpers (legacy + new-spec request modes)src/checks/*.ts— one file per readiness check; each exports aCheckDefinition. Registered inchecks/index.tssrc/report.ts— grading (A–F over decided checks; warn = half credit), terminal + JSON renderingtest/— vitest unit tests (pure functions only; probe logic is tested against live reference servers, see below)
Critical rules
- Verify spec details before implementing any probe. The scaffold's header names, method names, and error codes were written from summaries and are marked
TODO(verify). The sources of truth are the RC spec (https://blog.modelcontextprotocol.io/posts/2026-07-28-release-candidate/, the spec repo) and the SEPs behind each change: SEP-2567 (sessions removed), SEP-2575 (initialize removed), SEP-2243 (routing headers), SEP-2549 (caching metadata). A false "not ready" verdict against a well-known server destroys the credibility this project exists to build. - Test every check against reference servers before trusting it — at least one old-spec server (expected: fails) and one new-spec/RC server (expected: passes). The official TypeScript SDK betas (https://blog.modelcontextprotocol.io/posts/sdk-betas-2026-07-28/) can run both.
- Zero runtime dependencies. devDependencies only. This is a deliberate positioning choice for a probe tool.
- Checks must never throw for "server is broken" — that's a
failresult. Throwing is reserved for probe bugs (error) and unimplemented checks (NotImplementedError→todo). - Optional spec features are
warn, neverfail(e.g. cache metadata, deprecated-feature usage). - Registry scan publishes aggregates only — no named shame-list of servers/maintainers.
- Grades over partially-implemented check suites must be labeled partial (report.ts already does this).
- Never claim servers "break on July 28." The official position is "nothing breaks on July 28" — it's the spec-text publication date, not a switch-off; version negotiation continues and deprecated features live ≥12 months. All copy (README, CLI output, scan writeup) frames results as readiness/adoption, never breakage-on-a-date.
- Auth walls are not failures. 401/403 → preflight access
auth-required; checks reportskipped; exit code 2. The registry scan must publish denominators (open / auth-walled / unreachable / not-mcp) and compute percentages only over probeable servers.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 72 lines · 3,441 tokens per session scan A 2a074c36f1d4
mcp-spec-check CLAUDE.md is an instructions file published in the GitHub repository Roee-Tsur/mcp-spec-check (15 stars, last pushed 1mo ago), licensed MIT. It adds 3,441 tokens to every session, about $0.0172 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.