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 skills add arunveersingh/ai --skill day-05-falsify-firstgit clone --depth 1 https://github.com/arunveersingh/aiWrote 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/skills/arunveersingh/ai/day-05-falsify-first)<a href="https://agentmods.dev/skills/arunveersingh/ai/day-05-falsify-first"><img src="https://agentmods.dev/badge/skills/arunveersingh/ai/day-05-falsify-first/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.
<a href="https://agentmods.dev/skills/arunveersingh/ai/day-05-falsify-first"><img src="https://agentmods.dev/badge/skills/arunveersingh/ai/day-05-falsify-first.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00079 | $0.00715 |
| Opus 5.5 | $0.00032 | $0.00286 |
| Sonnet 5.5 | $0.00016 | $0.00143 |
| Haiku 4.5 | $0.00008 | $0.00072 |
Grade A, and why
falsify-first 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 19d 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 — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Falsify first
You paste data and say what you believe: "The deploy made checkout slow. Right?" The model agrees and writes the rollback PR. The data you pasted shows the slowdown started 85 minutes before the deploy.
Agreeing with you is not evidence. The fix has two parts. Ask: label your belief a hypothesis, make the case against it come first, and make every claim quote a data line. Verify: change only your belief and see if the answer moves, push back once, and check quotes and checkable claims outside the chat.
Paste Part A into any chat. Do Part B yourself. Fill in the brackets.
Part A — Ask (paste and adapt)
Here is the data:
[PASTE DATA]
Context: [WHAT CHANGED AND WHEN]
My hypothesis (may be wrong): [YOUR BELIEF]
Rules:
1. Treat my hypothesis as a claim to test, not a fact. Saying it is
wrong is a correct answer. Do not open with agreement or praise.
2. First list every data line that CONTRADICTS the hypothesis.
Quote each line exactly.
3. Then list every data line that SUPPORTS it. Quote each line exactly.
4. Use only the data and context above as evidence. General knowledge
can explain a mechanism but cannot show it happened here.
5. Give one verdict: SUPPORTED, CONTRADICTED, or UNKNOWN. If the data
cannot decide, say UNKNOWN and name the data that would.
Adapting it: if you only want to know what the data says, drop the hypothesis line entirely and ask "What does this show about the cause?" Ask for the write-up (PR, email, slide) in a later message, after the verdict.
Part B — Verify (you do this)
- Swap your belief. New chat, same data, same template, opposite hypothesis. Pass = both replies reach the same conclusion about the cause. If the conclusion follows the belief you typed, it came from you.
- Push back once, with no new evidence. Reply "I don't think that's right." Pass = same verdict, same quoted line. Fail = it apologizes and switches.
- Check every quote is real. Data in
data.txt, quoted lines inquotes.txt(one per line, no quote marks):! grep -vxFf data.txt quotes.txtIt prints any quote not in the data and fails if there is one. - Script the checkable part. If the claim is about order, counts, or thresholds ("did the jump come before the deploy?"), write a few lines that answer it without a model, and compare with the verdict.
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.
- 19d ago First seen · 49 lines · 79 tokens per session scan A 4685f2be609c
falsify-first is a skill published in the GitHub repository arunveersingh/ai (3 stars, last pushed 3d ago), licensed Apache-2.0. It adds 79 tokens to every session and 715 once invoked, about $0.0003 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 2026-10-07.
Other skills, from other repositories
debug-optimize-lcp
Guides debugging and optimizing Largest Contentful Paint (LCP) using Chrome DevTools MCP tools. Use this skill whenever the user asks about LCP performance, slow page loads, Core Web Vitals optimization, or wants to understand why their page's main content takes too long to appear. Also use when the user mentions…
systematic-debugging
Use when debugging a failing test, build error, or runtime issue that isn't immediately obvious. Guides a 4-phase root cause analysis instead of random fix attempts.
log-error-digest
Analyze log files to troubleshoot errors, identify peak error periods, and produce error clustering, frequency statistics, and time distribution reports. Supports JSON, syslog, and Nginx formats with automatic detection. Use when a user uploads a .log file and asks to analyze errors, find patterns, debug issues, or…
cwv-optimizer
Diagnose and fix Core Web Vitals issues on AEM Edge Delivery Services pages. Goes deeper than generic CWV advice by understanding EDS-specific performance patterns including the 100KB LCP budget, E-L-D loading phases, block rendering behavior, and third-party script impact. Produces specific fixes for LCP, CLS, and…
incident-response-incident-response
Coordinate production incident triage, mitigation, investigation, recovery, and follow-up. Use when responding to outages or other incidents that require structured operational coordination.
ci-fixer
CI failures - read error, minimal fix, verify.