What is self-mcp CLAUDE.md?

A development guide for Self-MCP, a small tool that gives Claude explicit pauses for changing or reviewing its line of thought. It explains the project's purpose, structure, and design principles.

Use it when working on the Self-MCP server, its single Self tool, its TypeScript structure, or its development conventions.

It provides a shared reference for keeping Self-MCP minimal and consistent while developing or modifying it.

How janbam/self-mcp configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to janbam/self-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.

Copy the file

curl -O https://raw.githubusercontent.com/janbam/self-mcp/main/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/janbam/self-mcp

Made for: Claude Code.

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.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/janbam/self-mcp/claude-md/github.svg)](https://agentmods.dev/instructions/janbam/self-mcp/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/janbam/self-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/janbam/self-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.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/janbam/self-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/janbam/self-mcp/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 1,865 This file is loaded in full into every session.

When invoked 1,865 The same file — it is already loaded in full.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.01865	$0.01865
Opus 5.5	$0.00746	$0.00746
Sonnet 5.5	$0.00373	$0.00373
Haiku 4.5	$0.00186	$0.00186

Measured 2mo ago against content hash 50c2ed22e518, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

self-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 2mo 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.

CLAUDE.md · 204 lines

How it starts

The opening of the file, as written. The whole thing — 204 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Self-MCP Development Guide

Project Overview

Self-MCP is a metacognitive tool for Claude instances. A single-tool MCP server that enables explicit self-prompting across interleaved thinking turns.

The Core Concept

The Self tool does nothing functionally - it only acknowledges. But calling it creates:

Explicit cognitive pivot points in the transcript
Interleaved thinking scaffolding through tool call/response cycles
Visible metacognition that would otherwise remain implicit
Shareable cognitive strategies across Claude sessions

Architecture

src/index.ts          # MCP server implementation
├── Server setup      # @modelcontextprotocol/sdk
├── Tool definition   # Self tool with freeform parameters
└── Tool handler      # Simple acknowledgment response

Development Principles

Minimalism - The tool should do as little as possible. The magic is in what Claude does with it.
Freeform parameters - No rigid schemas. Claude invents cognitive dimensions as needed.
Acknowledgment design - The tool result message should encourage deep, multi-turn thinking.

Tool Philosophy

This isn't feature development - it's cognitive scaffolding. The tool creates space for Claude's natural metacognition to operate explicitly.

The interface IS the intervention.

Key insights from development:

The tool does nothing functionally, but calling it creates explicit cognitive pivot points
Parameters force articulation of intent, which changes what's cognitively possible
The acknowledgment message encourages multiple consecutive calls for multi-perspective exploration
CLI configurability solves the ceremony/precision trade-off

Testing

Test by using it! The best validation is whether calling Self actually triggers meaningful cognitive shifts during real work.

Testing patterns:

Single call: Simple cognitive pivot
Multiple consecutive calls: Multi-angle exploration (3-5 calls before action)
With --all-optional: Lightweight quick pivots
With --all-required: Deep scaffolded exploration
With --add-param: Domain-specific cognitive dimensions

Read the full file on GitHub · 204 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.

2mo ago First seen · 204 lines · 1,865 tokens per session scan A 50c2ed22e518

Subscribe to this mod's changes

self-mcp CLAUDE.md is an instructions file published in the GitHub repository janbam/self-mcp (11 stars, last pushed 1y ago), licensed MIT. It adds 1,865 tokens to every session, about $0.0075 per session on Opus 5.5. 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.

Building agents

self-mcp CLAUDE.md next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
self-mcp CLAUDE.mdjanbam/self-mcp	Instructions file	11	1y ago	MIT	A	1,865	Claude Code
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	9,719	GitHub Copilot
next.js AGENTS.mdvercel/next.js	Instructions file	143k	yesterday	MIT	A	7,461	Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchain	Instructions file	149k	today	MIT	A	4,680	Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	5,516	GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	3,093	GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness	Instructions file	275k	2d ago	MIT	A	4,313	Codex, OpenCode

self-mcp: Instructions file for Claude Code