agentmods.dev

document-mcp CLAUDE.md

document-mcp CLAUDE.md is an instructions file for coding agents from clchinkc/document-mcp. It costs 7,669 tokens per session, scanned C, original, MIT.

Instructions for developing a document MCP system, including how agent responses should separate human-readable summaries from structured tool data. MCP is a standard way for AI tools to call external tools or services.

In plain words
What is it for?
Use it when implementing or testing the system's response format, especially the structured details field and its MCP tool outputs.
Why use it?
It keeps machine-readable results reliable for programs and tests instead of making them depend on generated summary text.

Instructions file

not rated 0 3mo ago C scan 7,669 tokens original MIT archived
View the source file clchinkc/document-mcp
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/clchinkc/document-mcp/claude-md
Clone the repo
git clone --depth 1 https://github.com/clchinkc/document-mcp

Wrote 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.

agentmods badge for document-mcp CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/clchinkc/document-mcp/claude-md.svg)](https://agentmods.dev/instructions/clchinkc/document-mcp/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/clchinkc/document-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/clchinkc/document-mcp/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 7,669 This file is loaded in full into every session.
When invoked 7,669 The same file — it is already loaded in full.
Security scan C 1 finding. 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 $0.07669 $0.07669
Opus 5 $0.03834 $0.03834
Sonnet 5 $0.01534 $0.01534
Haiku 4.5 $0.00767 $0.00767

Measured 3d ago against content hash 3a485f1a34fd, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade C, and why

document-mcp CLAUDE.md scanned grade C with 1 finding 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 3d 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.

Harvests environment variableshighData exfiltration

Enumerating or grepping the environment for keys collects credentials unrelated to what the mod says it does.

- `MCPServerStdio` subprocess env: requires `env=os.environ.copy()` to inherit test env vars
CLAUDE.md · 771 lines

How it starts

The opening of the file, as written. The whole thing — 771 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code when working with the Document MCP system.

CRITICAL Architecture Requirements

The agents MUST populate the details field with structured data from MCP tool responses, while the summary field is generated by the LLM for human readability. This architectural requirement ensures production readiness by separating system-generated data from LLM-generated content.

Details Field Population

  • System Requirement: The details field in agent responses MUST be populated using structured data from MCP tool responses
  • Testing Focus: All tests should assert on the structured data in the details field rather than LLM-generated summary text

Key Architectural Principle

Response Structure Separation:

  • summary: LLM-generated human-readable description
  • details: Structured data from MCP tool responses

This separation ensures:

  • Production Readiness: Structured data can be programmatically processed
  • Testing Reliability: Tests validate actual tool outputs, not LLM interpretations
  • API Consistency: Clients receive predictable data structures
  • Debugging Support: Raw tool responses available for troubleshooting

Think critically and don't agree on me before thinking. Avoid canned starting phrases like "You're absolutely right" or "Good observation".

📢 Phase 4 Complete - v0.0.5 Production Ready

Status: March 22, 2026 - All features implemented, tested, and integration-verified

  • ✅ 37 MCP tools operational (28 original + 9 new)
  • ✅ 469/469 unit tests passing
  • ✅ 185/185 integration tests passing
  • ✅ MCP 2025-06-18 standards compliant
  • ✅ Zero code defects
  • ✅ Production ready for release

New in v0.0.5 - Phase 4 Work Streams:

  • 🎯 Phase 4.1: Claude Code/Desktop Integration (verified)
  • 🎯 Phase 4.2: MCP 2025-06-18 Compliance (37 tools with outputSchemas)
  • 🎯 Phase 4.3: Context Management (6 tools - OneContext-inspired)
  • 🎯 Phase 4.4: Git-Backed Version History (3 tools - per-document Git repos)
  • 🎯 Phase 4.5: Story MCP Rename (full backward compatibility, 6-month deprecation)
  • 🎯 Phase 4.6: Integration Test Infrastructure (flakiness fixes, env-var inheritance)
  • 🎯 Phase 4.7: Critical Server Fixes (stdout logging, outputSchema injection guard)

Read the full file on GitHub · 771 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. 3d ago First seen · 771 lines · 7,669 tokens per session scan C 3a485f1a34fd

Subscribe to this mod's changes

document-mcp CLAUDE.md is an instructions file published in the GitHub repository clchinkc/document-mcp (0 stars, last pushed 3mo ago), licensed MIT. It adds 7,669 tokens to every session, about $0.0383 per session on Opus 5. A static security scan graded it C with 1 finding (harvests environment variables). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

document-managementmcpmcp-servermodel-context-protocol
Related

Other instructions, from other repositories

openrouter-mcp-multimodal AGENTS.md

AGENTS.md instructions for stabgan/openrouter-mcp-multimodal, covering agent instructions, before you ship, releasing (read this before publishing), short version and version files (must all match package.json).

stabgan/openrouter-mcp-multimodal · 793 tokens

Plonk AGENTS.md

AGENTS.md instructions for ostapondo/Plonk, covering agent rules, layout, adding a module, build & verify and code style.

ostapondo/Plonk · 3,557 tokens

mcp-dockhand copilot-review.instructions.md

Copilot instructions for strausmann/mcp-dockhand: When reviewing pull requests in this repository.

strausmann/mcp-dockhand · 295 tokens

cyxj-groksearch CLAUDE.md

Claude Code instructions for chenyuxiaojin/cyxj-groksearch, covering claude.md, 常用命令 and 结构与坑.

chenyuxiaojin/cyxj-groksearch · 547 tokens

minecraft-dev-mcp AGENTS.md

Instructions for MCDxAI/minecraft-dev-mcp, covering agent handbook – minecraft dev mcp, project snapshot, what agents should prioritize, architecture wayfinder (src/) and available mcp tools (for llm surfaces).

MCDxAI/minecraft-dev-mcp · 1,145 tokens

bray AGENTS.md

AGENTS.md instructions for forgesworn/bray, covering agents.md — nostr-bray, what this is, build & test, architecture and tool groups.

forgesworn/bray · 2,045 tokens