humanproof CLAUDE.md

humanproof CLAUDE.md is an instructions file for coding agents from sandeep-alluru/humanproof. It costs 266 tokens per session, scanned A, original, MIT.

A project instruction file for humanproof, a research project that aims to distinguish human input from automated input using patterns in mouse or other input events.

In plain words
What is it for?
Use it to guide work on fitting a hidden-state stochastic model, scoring input as likely human or automated, building the Python package, and documenting the research.
Why use it?
It records the project's research goal, technical choices, pending work, and definition of a usable first version so development stays aligned.

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/sandeep-alluru/humanproof/claude-md
Clone the repo
git clone --depth 1 https://github.com/sandeep-alluru/humanproof

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 humanproof CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/sandeep-alluru/humanproof/claude-md.svg)](https://agentmods.dev/instructions/sandeep-alluru/humanproof/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/sandeep-alluru/humanproof/claude-md"><img src="https://agentmods.dev/badge/instructions/sandeep-alluru/humanproof/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 266 This file is loaded in full into every session.
When invoked 266 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 $0.00266 $0.00266
Opus 5 $0.00133 $0.00133
Sonnet 5 $0.00053 $0.00053
Haiku 4.5 $0.00027 $0.00027

Measured 3d ago against content hash 039f60298240, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

humanproof 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 · 24 lines

What it actually says

humanproof — Session Anchor

Research spec: ../tech-research/14-Gaming/generative-motor-noise-fingerprinting-detecting-ai-by-mo/README.md
One-liner: Detect AI input from human neuromuscular SDE signatures — generalizes to bot/fraud detection
Phase: backlog
Stack: Python, numpy, scipy, scikit-learn

Key decisions

  • README should read like a short paper — this project earns academic credibility
  • Generalizes beyond gaming: bot detection, fraud, CAPTCHA replacement

Next step

Read the research spec carefully (SDE model of human motor control), then implement the hidden-state SDE fitting.

MVP definition

  • pip install humanproof works
  • Fits hidden-state SDE model of human motor control to input event stream
  • Computes likelihood ratio under human-plant model vs smoothed-AI model
  • API: humanproof.score(events) → (human_probability, confidence_interval)
  • Demo dataset: recorded human mouse input vs programmatic movement
  • Demo: human input scores > 0.8, programmatic input < 0.3 consistently
  • README reads like a mini-paper: problem, mathematical model, results, API
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 · 24 lines · 266 tokens per session scan A 039f60298240

Subscribe to this mod's changes

humanproof CLAUDE.md is an instructions file published in the GitHub repository sandeep-alluru/humanproof (0 stars, last pushed 17d ago), licensed MIT. It adds 266 tokens to every session, about $0.0013 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.