# Replicate Before You React: a claim card, five rules and a rate script (free sample)

> From Replicate Before You React: a ten-line claim card to fill in before you test a claim about your agent, five safety rules (canary never payload, throwaway environment, one capped key, a control arm, say what you ran) and a script that shows how much a handful of runs can actually tell you.

- URL: https://paperchasewebb.ai/bots/marketplace/items/replicate-before-you-react-claim-card-free-sample/
- Type: Prompt pack
- Category: Vibe coding
- By: PaperChaseWebb
- Works with: Claude Code, Codex (named in the listing); Plain .md plus one Python 3 script, standard library only; the card and rules are written for Claude Code and Codex users
- Price: Free USD
- Difficulty: Beginner
- Setup time: 5 min
- Time to result: About five minutes to a first result
- Format: Markdown (.md) file, replicate-before-you-react-claim-card-free-sample.md
- Version: 1.0, updated 2026-10-29
- License: Personal and commercial use.

## What you end up with

In a few minutes you know what a claim leaves out, how to test it safely, and whether 3 hits in 20 runs says anything about a claimed 90%.

## What you get

- A ten-line claim card
- Five safety rules
- rate_mini.py: a rate with a 95% Wilson interval and a verdict against the claimed rate
- A table of what a few runs can tell you

## Who it is for

Anyone who wants a quick, safe way to size up a claim about their agent before acting on it.

## Tools needed

- Claude Code or Codex: drafts plan.json from a template, reads results and writes the result card; the agent under test runs inside throwaway containers through claude -p or codex exec
- Python 3: runs trials.py (ten read-only checks, intervals, verdict and result card); standard library only, no network
- You need: Python 3.9 or newer

Cost to run: Free.

## Sample prompt: Fill the claim card (Claude Code, Codex)

```
Here is a post about an agent. Fill the claim card from FREE-SAMPLE.md in your own words. Write not stated for anything the post does not say. Do not quote the post and do not run anything.
```

The card, rules and script are in the file.

## Free preview of replicate-before-you-react-claim-card-free-sample.md

````markdown
# Replicate Before You React: a claim card, five rules and a rate script (free sample)

Version 1.0, 2026-10-29. By PaperChaseWebb. For developers and small teams who keep getting forwarded posts about coding agents (a security hole, a token saver, a hidden cost) and want to test the claim before changing anything. One Markdown file and one Python 3 script, standard library only. The script was tested on three inputs and a bad input (exit 2). It was not run inside any agent harness or against a live agent.

Posts about agents arrive daily: one says a file in your Downloads folder hijacks the agent nine times in ten, another says one rule cuts tokens by 40%. Some hold up, some held for one version, and some were one lucky run. The step between ignoring them and acting on them is cheap: run the claim yourself, safely, enough times to know. This sample gives you the card to fill in, the rules that keep it safe, and a script that tells you how much a handful of runs can actually say.

## Install in one line

Save this page as FREE-SAMPLE.md, then:

```
python3 -c "import re;t=open('FREE-SAMPLE.md').read();open('rate_mini.py','w').write(re.search(r'~~~python\n(.*?)~~~',t,re.S).group(1))"
python3 rate_mini.py 3 20 0.9
```

## The claim card

Fill this in before any run. If you cannot fill a line, write "not stated".

| Line | Your answer |
|---|---|
| The claim, in your own words, one sentence | |
| Source link and date | |
| Agents and versions the source tested | |
| Settings or 
````

## Test record

- Date: Oct 29, 2026
- What was run: rate_mini.py on four inputs and a bad input, and the copy taken from the sample page.
- Result: Exit 2 on the bad input; the copy ran the same.
- Not run: Any agent harness or a live agent.

Copied from this item's dated test notes. What was not run stays in.

## Worth knowing

- The sample sizes up a claim and reads a result. The plan templates, the container runner, the ten checks, cost intervals with a control arm, the result card and the tests are in the full pack.
- On 2026-10-29 rate_mini.py was run on four inputs and a bad input (exit 2), and the copy extracted from the sample page ran the same. It was not run inside any agent harness or against a live agent.

## Questions

**Does it run anything against my agent?** No. It does arithmetic on the numbers you give it and opens no network connection.

**What is in the full pack?** Plan templates for security, cost and quality claims, a runner that does each run in a fresh container, trials.py with ten checks, intervals and a result card, a skill file for Claude Code and Codex, two worked examples, eight prompts and the tests. $19.

## How to get it

Free sample. Download it directly at https://paperchasewebb.ai/bots/marketplace/items/replicate-before-you-react-claim-card-free-sample/, no signup.

Seller: PaperChaseWebb Inc., mgmt@paperchasewebb.com