agentmods.dev

OmniGlyph CLAUDE.md

A repository instruction document for OmniGlyph, a proxy that compresses large language-model request content into small one-bit PNG pages when that reduces cost and passes a reading test. It also records how to install, build, test, run, and type-check the project.

In plain words
What is it for?
Use it when working on OmniGlyph to understand its source folders, run the build and tests, check types, start development servers, and follow project-specific rules.
Why use it?
It gives coding agents the project's architecture, conventions, commands, and required checks in one place. This reduces guesswork when changing a repository with specialized compression and billing logic.

Instructions file

104 7d ago A scan 987 tokens original MIT
View the source file diegosouzapw/OmniGlyph
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/diegosouzapw/omniglyph/claude-md
Clone the repo
git clone --depth 1 https://github.com/diegosouzapw/OmniGlyph
Per session 987 This file is loaded in full into every session.
When invoked 987 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 $0.00987 $0.00987
Opus 5 $0.00494 $0.00494
Sonnet 5 $0.00197 $0.00197
Haiku 4.5 $0.00099 $0.00099

Measured yesterday against content hash d02e9ddaa44f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

OmniGlyph 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 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.

CLAUDE.md · 74 lines

How it starts

The opening of the file, as written. The whole thing — 74 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 (claude.ai/code) when working with code in this repository.

Quick Start

pnpm install               # install deps
pnpm run build             # esbuild bundle → dist/ (Node + Workers entrypoints)
pnpm run typecheck         # tsc --noEmit (must be clean)
pnpm test                  # vitest run — full suite (can take >2 min; run in background)
pnpm run dev:node          # local proxy at http://127.0.0.1:47821 (tsx watch)
pnpm run dev:worker        # Cloudflare Workers host (wrangler dev)

Focused test runs (preferred while iterating): npx vitest run tests/<file>.test.ts.

Project at a Glance

Context-as-image compression proxy: rewrites bulky LLM request blocks (system prompt, tool docs, old history, large tool outputs) into dense 1-bit PNG pages with exact per-provider billing math. Only converts when the math wins, and only for models that passed the reading benchmark (fail-closed gate).

layer where what
Core src/core/ transform pipeline, exact billing (anthropic-vision.ts, gemini-model-profiles.ts), render cache, proxy logic
Render src/core/render* + assets/ 1-bit 5×8 glyph atlas (Spleen/Unifont-derived), reflow, PNG pages
Hosts src/node.ts · src/worker.ts local Node server (dashboard + events) · Cloudflare Workers
Public API src/core/index.ts billing math re-exported at package root (consumed by OmniRoute)
Benchmarks benchmarks/ billing-sweep + density-frontier harnesses; JSONL receipts in */results/
Tests tests/ vitest; includes docs-integrity (link rot + rebrand guard)

Key Conventions

  • Strict TDD: write the failing test first, watch it fail for the right reason, then the minimal implementation. No production code without a test.
  • Measurement before claims: any number in docs/README must have a receipt in benchmarks/*/results/. Measured decisions (see docs/benchmarks/BENCHMARKS.md) are not re-litigated without new data.
  • TypeScript strict, ESM only (.js import suffixes in TS source).
  • Comments state constraints the code can't show — not narration.
  • User-facing brand strings are OmniGlyph; binary/env prefix is omniglyph / OMNIGLYPH_*; data in ~/.omniglyph/, config in ~/.config/omniglyph/.

Read the full file on GitHub · 74 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 · 74 lines · 987 tokens per session scan A d02e9ddaa44f

Subscribe to this mod's changes

OmniGlyph CLAUDE.md is an instructions file published in the GitHub repository diegosouzapw/OmniGlyph (104 stars, last pushed 7d ago), licensed MIT. It adds 987 tokens to every session, about $0.0049 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 2026-08-30.

Data and AIMemory and contextTestinganthropicclaudeclaude-codecompressioncontext-windowgemini
Related

Other instructions, from other repositories

tokenfold AGENTS.md

Instructions for snchimata/tokenfold, covering repository guidelines, working principles, repository map, implementation conventions and validation.

snchimata/tokenfold · 894 tokens

distil CLAUDE.md

Claude Code instructions for munhq/distil, covering distil — context optimization for llm agents, what this is, consumption modes, architecture and layers (src/layers/).

munhq/distil · 1,447 tokens

densely CLAUDE.md

Instructions for alibaizhanov/densely: Behavioral guidelines to reduce common LLM coding mistakes. Merge with project-specific instructions as needed.

alibaizhanov/densely · 711 tokens

DeepSpeed AGENTS.md

AGENTS.md instructions for deepspeedai/DeepSpeed, covering deepspeed project rules, commit & ci requirements, code change discipline, tool caveats and edit tool auto-formatter.

deepspeedai/DeepSpeed · 455 tokens

AionUi AGENTS.md

Instructions for iOfficeAI/AionUi, covering aionui - project guide, code conventions, file & directory structure, naming and ui library & icons.

iOfficeAI/AionUi · 1,663 tokens

botmux CLAUDE.md

Instructions for deepcoldy/botmux, covering botmux, 构建 & 运行, bun 开发链路, worktree 的 nodemodules：共享还是独立（改前必读） and 编译态（单文件二进制）注意.

deepcoldy/botmux · 4,000 tokens