Borrowing it
Nothing to install: this file belongs to YGao2005/scholar-feed-mcp. 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/YGao2005/scholar-feed-mcp/main/CLAUDE.mdgit clone --depth 1 https://github.com/YGao2005/scholar-feed-mcpWrote 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/ygao2005/scholar-feed-mcp/claude-md)<a href="https://agentmods.dev/instructions/ygao2005/scholar-feed-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/ygao2005/scholar-feed-mcp/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/ygao2005/scholar-feed-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/ygao2005/scholar-feed-mcp/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 | $0.02034 | $0.02034 |
| Opus 5 | $0.01017 | $0.01017 |
| Sonnet 5 | $0.00407 | $0.00407 |
| Haiku 4.5 | $0.00203 | $0.00203 |
Grade A, and why
scholar-feed-mcp 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 5d 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 — 126 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Scholar Feed MCP — Project Instructions
The official MCP server for Scholar Feed (600k+ CS/AI/ML papers). Published to npm as
scholar-feed-mcp and hosted remotely at https://mcp.scholarfeed.org/mcp.
Deep dev/release detail lives in
CONTRIBUTING.mdand.local-docs/— this file is the load-bearing guardrails. When they conflict, fix the drift.This repo is PUBLIC. Internal specs, handoffs, runbooks, and launch/distribution material go in
.local-docs/(gitignored), never in a tracked path. Only user-facing docs —README.md,CONTRIBUTING.md— are published.
What this repo is (and isn't)
- A thin MCP client. It exposes tools (search, citations, lineage, library, watches, …) and
forwards to the real backend, which is a SEPARATE repo (
../scholar-feed, FastAPI on Heroku- Supabase). Client tools routinely ship ahead of the backend routes they call — a tool can exist here before its endpoint is live in prod.
- Do not put business logic, DB access, or scoring here. If a tool needs new data, the change
is usually a backend serializer/route in
../scholar-feed, then a thin tool here.
Architecture — three transports, one tool set
| Entry | File | Used by |
|---|---|---|
| stdio | src/index.ts (the npm bin) |
local npx scholar-feed-mcp, Claude Desktop/Code |
| HTTP (Express) | src/server.ts → src/server-http.ts |
the remote endpoint on Vercel (mcp.scholarfeed.org) |
| Cloudflare Worker | src/worker.ts |
edge deploy (npm run worker:deploy) |
Tools live in src/tools/ (one file per tool, registered in src/tools/index.ts).
src/client.ts is the backend API wrapper (auth + error handling).
🔴 Hard rules
- Publishing is AUTOMATED — never run
npm publishby hand.publish.ymlpublishes on a pushedv*tag via npm OIDC trusted-publishing (no token, with provenance). Release frommain:npm version <patch|minor|major>thengit push --follow-tags— the tag push IS the publish. (Assistant may prepare the bump+tag; the human pushes.) SeeCONTRIBUTING.md→ Releasing. - All logging to stderr (
console.error).console.log()corrupts the JSON-RPC stdio transport and silently breaks the client. - Keep the package self-contained:
dependenciesmust stay{}. tsup bundles everything (deps → devDeps) so cold-startnpxis ~1 pkg / 0 transitive installs. Adding a runtime dep re-introduces the cold-start timeout that shipped as a silent "failed to connect". - The remote endpoint is
src/worker.ts(Cloudflare), not the Express server. The historical trap here was Vercel zero-config auto-detecting the stdio bin and 500'ing the endpoint for two days, whichvercel.jsonexisted to prevent. Vercel is fully gone as of 2026-07-30: the account was cancelled,vercel.jsonis deleted, and the GitHub integration is uninstalled (it had been red-failing every PR with "deployment blocked"). If aVercelcheck ever reappears on a PR, the app got reinstalled — uninstall it rather than re-adding config.src/server.ts/src/server-http.tsstill backnpm run dev:httpfor local HTTP testing — don't delete them, but don't expect them to serve prod.SF_MCP_ALLOWED_HOSTSis required for any non-loopback deploy: unset = loopback-only, which rejects every real request. Never setSF_API_KEYon the remote surface (cross-tenant leak — each request derives its own key from the Authorization header).
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.
- 5d ago Changed · +26 lines · +451 tokens per session 9caa869616f3
- 10d ago First seen · 100 lines · 1,583 tokens per session scan A dbc6623d1cc3
scholar-feed-mcp CLAUDE.md is an instructions file published in the GitHub repository YGao2005/scholar-feed-mcp (10 stars, last pushed yesterday), licensed MIT. It adds 2,034 tokens to every session, about $0.0102 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 2026-08-31.
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).
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.
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.