What is llm-wiki-loop CLAUDE.md?

Repository instructions for a research wiki that records claims, their evidence, current status, and later results. bioRxiv-style source files are kept separate from summaries, while outcomes are added as a permanent history.

Use it when extracting facts from sources, maintaining a research wiki, recording evidence, or tracking whether claims hold up after testing.

It prevents source material from being silently rewritten to fit a summary and keeps each claim traceable over time.

How helicerat/llm-wiki-loop configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to helicerat/llm-wiki-loop. 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/helicerat/llm-wiki-loop/main/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/helicerat/llm-wiki-loop

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/helicerat/llm-wiki-loop/claude-md/github.svg)](https://agentmods.dev/instructions/helicerat/llm-wiki-loop/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/helicerat/llm-wiki-loop/claude-md"><img src="https://agentmods.dev/badge/instructions/helicerat/llm-wiki-loop/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/helicerat/llm-wiki-loop/claude-md"><img src="https://agentmods.dev/badge/instructions/helicerat/llm-wiki-loop/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 983 This file is loaded in full into every session.

When invoked 983 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	$0.00983	$0.00983
Opus 5.5	$0.00393	$0.00393
Sonnet 5.5	$0.00197	$0.00197
Haiku 4.5	$0.00098	$0.00098

Measured 2mo ago against content hash a4b89b8b43d8, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

llm-wiki-loop 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 · 107 lines

How it starts

The opening of the file, as written. The whole thing — 107 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Schema

This file is the contract. It holds only rules that cause damage if forgotten mid-session, because this is the file that survives context compaction. Everything else lives in wiki/index.md, which may compress away and be re-read on demand.

Layers

raw/ — sources. Read them, never write to them, never move a file out of wiki/ back into raw/. Ingestion is one-way. This boundary is a directory, not a request, because the failure it prevents — silently rewriting a source to match a summary — is invisible once it happens.

wiki/ — yours entirely. Create pages, update them, maintain cross-references.

outcomes/ — append-only. What shipped, and what happened afterward. Never edit a past entry; correct it with a new one.

The claim contract

Every claim page carries:

claim: one sentence, falsifiable
status: candidate | rule | retired
n: 0
effect: observed magnitude, with units
form: timeless | snapshot | pointer
as_of: YYYY-MM-DD
provenance: extracted | inferred | ambiguous
outcomes: []

n is not editable by hand. It is incremented only by wiki-measure, only from an entry in outcomes/. A claim whose n cannot be reconstructed from the outcome log is a lint error, not a rounding difference.

State transitions, mechanical, no judgement:

candidate → rule at n >= 3
rule → retired when the three most recent outcomes contradict it
nothing is deleted, ever. Retired claims keep their history

Only status: rule may block a draft. Candidates are reported and never enforced. This is the whole discipline: an untested observation does not get to overrule you.

Three legal forms for a fact

timeless — does not decay. No stamp needed
snapshot — dated observation. Requires as_of
pointer — names where the truth actually lives. Requires a resolvable target

A present-tense quantitative claim about something volatile, carrying neither as_of nor pointer form, is a lint error. Anything inside a dated container is exempt structurally and needs no annotation.

Read the full file on GitHub · 107 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 · 107 lines · 983 tokens per session scan A a4b89b8b43d8

Subscribe to this mod's changes

llm-wiki-loop CLAUDE.md is an instructions file published in the GitHub repository helicerat/llm-wiki-loop (6 stars, last pushed 3mo ago), licensed MIT. It adds 983 tokens to every session, about $0.0039 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.

Research

llm-wiki-loop CLAUDE.md next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
llm-wiki-loop CLAUDE.mdhelicerat/llm-wiki-loop	Instructions file	6	3mo ago	MIT	A	983	Claude Code
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	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	yesterday	MIT	A	4,680	Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	5,516	GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	3,093	GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness	Instructions file	274k	2d ago	MIT	A	4,313	Codex, OpenCode

llm-wiki-loop: Instructions file for Claude Code