yallmap CLAUDE.md

yallmap CLAUDE.md is an instructions file for coding agents from kckempf/yallmap. It costs 340 tokens per session, scanned A, original, MIT.

Project instructions for yallmap, a TypeScript gateway that forwards requests between Claude Code and Anthropic's API and records tracing data for inspection.

In plain words
What is it for?
Use them when developing the gateway's message route, request routing, streaming, deployment setup, and Langfuse-based observability.
Why use it?
They preserve the project's agreed architecture and warn about details that can easily break, such as streaming responses, authentication headers, and tool-use fields.

Instructions file

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/kckempf/yallmap/claude-md
Clone the repo
git clone --depth 1 https://github.com/kckempf/yallmap

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 yallmap CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/kckempf/yallmap/claude-md.svg)](https://agentmods.dev/instructions/kckempf/yallmap/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/kckempf/yallmap/claude-md"><img src="https://agentmods.dev/badge/instructions/kckempf/yallmap/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 340 This file is loaded in full into every session.
When invoked 340 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 $0.00340 $0.00340
Opus 5 $0.00170 $0.00170
Sonnet 5 $0.00068 $0.00068
Haiku 4.5 $0.00034 $0.00034

Measured 5d ago against content hash 4c910693a5c3, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

yallmap 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.

CLAUDE.md · 35 lines

What it actually says

LLM Gateway Project Context

What this is

Anthropic-native LLM gateway in TypeScript. Primary client is Claude Code itself via ANTHROPIC_BASE_URL. Emits OTel traces to self-hosted Langfuse.

Architectural decisions already made

  • Anthropic API first (not OpenAI), because Claude Code is the primary workload
  • Routing policies as TypeScript functions, not YAML/CEL/config
  • OTel gen_ai.* semantic conventions for trace emission
  • Langfuse as the observability backend (self-hosted)
  • CDK constructs for deployment (extends my existing cdk-dms-replication work)

v0.1 scope (this session)

  • POST /v1/messages route only
  • Transparent pass-through to api.anthropic.com
  • Preserve all headers (x-api-key, anthropic-version, anthropic-beta)
  • SSE streaming piped through — never buffer
  • OTel span emission to Langfuse

Gotchas to design for

  • Streaming: pipe upstream response directly, never .on('data') and collect
  • Tool-use IDs: preserve every field, don't strip unknown ones
  • Auth headers: don't rewrite x-api-key unless upstream expects different scheme

Differentiation hooks (later, not v0.1)

  • Hybrid local/cloud routing (M5 Max Ollama + Anthropic)
  • Code-based policies via TypeScript predicates
  • CDK constructs for AWS deployment

What I'm NOT building

  • OpenAI compatibility (defer)
  • Routing logic (defer)
  • Caching (defer)
  • Anything beyond transparent proxy + observability for v0.1
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. 5d ago First seen · 35 lines · 340 tokens per session scan A 4c910693a5c3

Subscribe to this mod's changes

yallmap CLAUDE.md is an instructions file published in the GitHub repository kckempf/yallmap (2 stars, last pushed yesterday), licensed MIT. It adds 340 tokens to every session, about $0.0017 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.