htmldeck CLAUDE.md

htmldeck CLAUDE.md is an instructions file for coding agents from uchimata2/htmldeck. It costs 3,923 tokens per session, scanned A, original, MIT.

Project instructions describing the working rules, loading order, scope, and conventions for the htmldeck codebase.

In plain words
What is it for?
They are for guiding work in the htmldeck folder, including checking loaded files, following the project’s instruction tiers, and respecting its documented boundaries.
Why use it?
They give the agent context about which instructions apply automatically and how to avoid relying on claims that the system does not enforce.

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/uchimata2/htmldeck/claude-md
Clone the repo
git clone --depth 1 https://github.com/uchimata2/htmldeck

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 htmldeck CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/uchimata2/htmldeck/claude-md.svg)](https://agentmods.dev/instructions/uchimata2/htmldeck/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/uchimata2/htmldeck/claude-md"><img src="https://agentmods.dev/badge/instructions/uchimata2/htmldeck/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 3,923 This file is loaded in full into every session.
When invoked 3,923 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 $0.03923 $0.03923
Opus 5 $0.01962 $0.01962
Sonnet 5 $0.00785 $0.00785
Haiku 4.5 $0.00392 $0.00392

Measured 3d ago against content hash 3c9857a28cea, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

htmldeck 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 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.

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 · 220 lines

How it starts

The opening of the file, as written. The whole thing — 220 lines — stays where its author put it; the contents beside it link to each section on GitHub.

htmldeck — working conventions

Read this before doing anything in this folder.

What loads every turn, and what bounds it

Tier 1 is what the harness loads without being asked — a property of the tree, not a list anyone maintains. Establish it by observation and never from a file's claim about itself: a document that asserts a load discipline the harness does not implement is worse than one over budget, because the claim cannot be falsified and content keeps being written on the strength of it. Tier 2 is what a skill or workflow document pulls in when work of a kind starts. Tier 3 is what tier 2 loads one at a time, for the branch actually taken.

Observed 2026-08-13: a session here receives three files unasked — the owner's global preferences, this file, and the memory index. Method: read what the session was given before its first tool call, then confirm the tree holds no second CLAUDE.md. This repository owns one of the three; the other two are the owner's and outside this bound, and the memory index is CE-10's. The plugin also writes into an adopter's tier 1, through skills/htmldeck/SKILL.md's description block and nothing else.

The bound: this file stays smaller than the smallest document it defers todocs/BRIEF.md, docs/PUBLISHING.md, tasks/TASK-WORKFLOW.md, tasks/README.md, .taskmd/config.md. Those five are the tier-2 set and the comparison is against tier 2, which is what makes the list closed rather than every document this file links to. A tier-3 document — one opened for a single question, never to start work of a kind — is not a term: docs/RELEASE-PHASES.md was never in the list and docs/RELEASE-HISTORY.md is not either. Otherwise splitting content out could never satisfy the bound, since each split makes a new document smaller than what it was cut from. Both terms are counted from the tree and no constant is written anywhere, because a number and the arithmetic that justified it have to be edited together and the number wins by staying put. The inequality says something: once the file you pay for on every turn costs more than any single document you open on demand, the split has inverted. Measure both terms with

Read the full file on GitHub · 220 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 · 220 lines · 3,923 tokens per session scan A 3c9857a28cea

Subscribe to this mod's changes

htmldeck CLAUDE.md is an instructions file published in the GitHub repository uchimata2/htmldeck (1 stars, last pushed 4d ago), licensed MIT. It adds 3,923 tokens to every session, about $0.0196 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.