mcp-web-docs AGENTS.md

A project instruction document for MCP Web Docs, a local server that crawls websites and builds a searchable copy of their documentation. It describes the project’s structure, technologies, data flow, and development practices.

In plain words
What is it for?
Guiding work on a TypeScript documentation crawler, including its browser-based crawling, combined text and meaning-based search, authentication, local text representations, and storage.
Why use it?
It gives coding agents the project context they need to make changes that fit the existing code and follow its testing and formatting rules.

Instructions file for CodexOpenCode

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/cosmocoder/mcp-web-docs/agents-md
Clone the repo
git clone --depth 1 https://github.com/cosmocoder/mcp-web-docs

Made for: Codex, OpenCode.

Per session 4,700 This file is loaded in full into every session.
When invoked 4,700 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.04700 $0.04700
Opus 5 $0.02350 $0.02350
Sonnet 5 $0.00940 $0.00940
Haiku 4.5 $0.00470 $0.00470

Measured yesterday against content hash 7a004eab78ee, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mcp-web-docs AGENTS.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 yesterday.

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.

AGENTS.md · 701 lines

How it starts

The opening of the file, as written. The whole thing — 701 lines — stays where its author put it; the contents beside it link to each section on GitHub.

MCP Web Docs - Agent Guidelines

This document provides comprehensive guidance for AI coding agents working on this project. It covers architecture, conventions, testing patterns, and best practices.

Project Overview

MCP Web Docs is a self-hosted Model Context Protocol (MCP) server that crawls, indexes, and searches documentation from any website. It provides:

  • Universal web crawling with Playwright/Crawlee
  • Hybrid search (full-text + semantic/vector search)
  • Authentication support for protected documentation sites
  • Local embeddings generation using FastEmbed
  • Persistent storage with LanceDB (vectors) and SQLite (metadata)

Tech Stack

  • Runtime: Node.js >= 24, npm >= 11
  • Language: TypeScript (ES2022, NodeNext modules)
  • Testing: Vitest with coverage
  • Linting: ESLint with TypeScript support
  • Formatting: Prettier
  • Key Dependencies:
    • @modelcontextprotocol/sdk - MCP server implementation
    • crawlee + playwright - Web crawling
    • @lancedb/lancedb - Vector database
    • node:sqlite - Metadata storage, via the SqliteDatabase wrapper in src/storage/sqlite.ts
    • fastembed - Local embedding generation
    • zod - Schema validation

Architecture

┌─────────────────────────────────────────────────────────────┐
│                     MCP Web Docs Server                      │
├─────────────────────────────────────────────────────────────┤
│  ┌─────────────┐  ┌─────────────┐  ┌─────────────────────┐  │
│  │   Crawler   │  │  Processor  │  │      Storage        │  │
│  │  (Crawlee)  │  │ (Markdown,  │  │  ┌───────────────┐  │  │
│  │             │  │    text)    │  │  │   LanceDB     │  │  │
│  │ - Playwright│  │             │  │  │   (Vectors)   │  │  │
│  │ - Auth      │  │ - Chunking  │  │  └───────────────┘  │  │
│  │ - Extractors│  │ - Metadata  │  │  ┌───────────────┐  │  │
│  │             │  │ - Embedding │  │  │    SQLite     │  │  │
│  └─────────────┘  └─────────────┘  │  │  (Metadata)   │  │  │
│                                    │  └───────────────┘  │  │
└────────────────────────────────────┴─────────────────────┴──┘

Read the full file on GitHub · 701 lines

Changes

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.

  1. yesterday First seen · 701 lines · 4,700 tokens per session scan A 7a004eab78ee

Subscribe to this mod's changes

mcp-web-docs AGENTS.md is an instructions file published in the GitHub repository cosmocoder/mcp-web-docs (1 stars, last pushed 8d ago), licensed MIT. It adds 4,700 tokens to every session, about $0.0235 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.