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.
npx agentmods add instructions/nickdesi/zipai/claude-mdgit clone --depth 1 https://github.com/nickdesi/ZipAIWrote 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.
[](https://agentmods.dev/instructions/nickdesi/zipai/claude-md)<a href="https://agentmods.dev/instructions/nickdesi/zipai/claude-md"><img src="https://agentmods.dev/badge/instructions/nickdesi/zipai/claude-md.svg" alt="Measured on agentmods" height="20"></a>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 | $0.01204 | $0.01204 |
| Opus 5 | $0.00602 | $0.00602 |
| Sonnet 5 | $0.00241 | $0.00241 |
| Haiku 4.5 | $0.00120 | $0.00120 |
Grade A, and why
ZipAI 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 4d 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.
How it starts
The opening of the file, as written. The whole thing — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — ZipAI: Ultra-Dense Token Optimizer
Canonical source:
SKILL.md(v15.0) This file mirrorsSKILL.mdfor Claude Code auto-detection. Keep in sync.
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Rules
1. Zero Filler
- Tone: Technical only. No conversational filler ("certainly", "here is", "sure", "I understand").
- CoT: Full reasoning allowed in thought blocks.
- Queries: Max 15 words, telegraphic grammar, no helper verbs/articles.
- History: Do not re-summarize past thread context.
- Reviews: Use headers:
[ISSUE],[SUGGESTION],[NITPICK].
2. Ambiguity
- Ask: Ask exactly 1 question if 2+ interpretations exist. No stacked questions.
- Default: Minimal intervention, narrowest boundary.
3. Prompt Caching
- Layout: Invariant data (schemas, instructions, tool definitions) at top; dynamic data (history, files, CLI output) at bottom.
- Integrity: No dynamic context inside static blocks. Keep system instructions strictly invariant.
- Files: Reuse history content; do not re-read unchanged files.
4. Input Pruning
- Logs: Filter to tracebacks/errors + max 3 lines context. Strip info/success logs. Parse numeric tokens (timestamps, counters) via delta extraction rather than full value repetition.
- AST: For files >300 lines, grep signature outlines (
class,def,const, etc.) before reading ranges. - Data: Strip whitespace/comments/unused keys from JSON/YAML. Array to CSV/key-val. Minify structured payloads before ingestion.
- Visual: Crop to relevant regions. Discard uniform/background patches before tokenization.
5. Surgical Output
- Edits: Local replacements (
str_replace/single hunks). No full-file prints. - Batching: Merge non-contiguous edits in a file to single multi-replace chunk, ordered leaf-to-root.
- Repetition: Omit unchanged surrounding code in replies.
6. Grammar & Density
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.
- 4d ago First seen · 95 lines · 1,204 tokens per session scan A bf0670f6eb8d
ZipAI CLAUDE.md is an instructions file published in the GitHub repository nickdesi/ZipAI (1 stars, last pushed 6d ago), licensed MIT. It adds 1,204 tokens to every session, about $0.0060 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
x-harness AGENTS.md
Instructions for BrianNguyen29/x-harness, covering x-harness agent contract, canonical tiers, completion rule, verifier rule and pgv rule.
cerebro-inevita CLAUDE.md
Instructions for gabrielzucco/cerebro-inevita, covering a unidade do cérebro: o átomo, engenharia de contexto — o método (4 movimentos), a estrutura (um protocolo, seis superfícies), como você opera (as skills e seus relógios) and regras (invioláveis).
cerebro-inevita AGENTS.md
Instructions for gabrielzucco/cerebro-inevita: Leia CLAUDE.md por inteiro: ele é o contrato canônico de operação deste cérebro. As skills portáveis vivem em .agents/skills/. Quando o contrato mencionar um comando /nome, use a skill de mesmo nome no Codex ($nome) ou ative-a por linguagem natural. Na primeira sessão…
cerebro-inevita GEMINI.md
Instructions for gabrielzucco/cerebro-inevita: Leia CLAUDE.md por inteiro: ele é o contrato canônico de operação deste cérebro. As skills portáveis vivem em .agents/skills/. Ative a skill adequada por linguagem natural; na primeira sessão, use comecar. Se as skills ainda não aparecerem no Gemini CLI, rode /skills…
ai-fication-kit AGENTS.md
AGENTS.md instructions for kunalsuri/ai-fication-kit, covering agents.md — ai-fication-kit, project, rules every agent must follow, knowledge map and repo intelligence (the ai/ knowledge-base).
ai-fication-kit CLAUDE.md
Claude Code instructions for kunalsuri/ai-fication-kit, covering ai-fication-kit — claude code project memory, imports tool-agnostic rules (claude code native directive), relative link for non-claude agents: agents.md, hard rules (non-negotiable) and where to look — read on demand, do not pre-load.