smell-the-numbers

smell-the-numbers is a skill for Claude Code from mikestangdevs/craft-skills. It costs 165 tokens per session (1,220 once invoked), scanned A, original, MIT.

A checking process for numbers produced by code, such as reports, charts, calculations, dashboards, and metrics. It investigates values that violate physical limits, business expectations, or their own history.

In plain words
What is it for?
Use it when output changes sharply, contains suspicious values such as all zeros or negative zero, or conflicts with expected real-world behavior.
Why use it?
It catches results that look plausible because the code ran successfully but are actually wrong.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

Part of the craft-skills plugin — 17 skills shipped together

Good fit Use it when output changes sharply, contains suspicious values such as all zeros or negative zero, or conflicts with expected real-world behavior.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/mikestangdevs/craft-skills/smell-the-numbers
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.

Any agent
npx skills add mikestangdevs/craft-skills --skill smell-the-numbers
Clone the repo
git clone --depth 1 https://github.com/mikestangdevs/craft-skills

Made for: Claude Code.

Or install craft-skills, the plugin that ships this one along with the rest of its 17 skills.

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 smell-the-numbers

README.md
[![agentmods](https://agentmods.dev/badge/skills/mikestangdevs/craft-skills/smell-the-numbers/github.svg)](https://agentmods.dev/skills/mikestangdevs/craft-skills/smell-the-numbers)
Your own site
<a href="https://agentmods.dev/skills/mikestangdevs/craft-skills/smell-the-numbers"><img src="https://agentmods.dev/badge/skills/mikestangdevs/craft-skills/smell-the-numbers/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.

agentmods 80×15 button for smell-the-numbers

Your own site · 80×15
<a href="https://agentmods.dev/skills/mikestangdevs/craft-skills/smell-the-numbers"><img src="https://agentmods.dev/badge/skills/mikestangdevs/craft-skills/smell-the-numbers.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 165 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,220 The whole file, excluding the scripts and references it only reads on demand.
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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00165 $0.01220
Opus 5 $0.00082 $0.00610
Sonnet 5 $0.00033 $0.00244
Haiku 4.5 $0.00016 $0.00122

Measured 11d ago against content hash 1f0f67411ec6, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

smell-the-numbers 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 11d 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.

skills/rigor/smell-the-numbers/SKILL.md · 72 lines

How it starts

The opening of the file, as written. The whole thing — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Smell the Numbers

The failure mode this fixes

The code runs, the page renders, the report generates — and the numbers are nonsense. A line carrying 1858% of its rated capacity. A formula panel showing all zeros. A -0 in a financial summary. A KPI that moved 40x between runs of "the same" scenario. Agents (and tired humans) accept these because the mechanism worked: no exception, no red test, output produced. "It runs" gets mistaken for "it's right."

The worse version is the cosmetic fix: clamp the value, hide the row, round the -0, smooth the curve — make the output look plausible while the defect underneath keeps feeding everything else. That converts a visible bug into an invisible one.

This skill installs the domain-expert reflex: read the output the way someone who knows the territory would, flag everything that contradicts physical limits, business reality, or its own history — and treat each flag as a root-cause investigation, not a display problem.

When to Use This Skill

  • Output just changed significantly after a refactor, migration, or "no-behavior-change" cleanup
  • A result is surprising in either direction — catastrophic where success was expected, or suspiciously perfect
  • You see sentinel smells: zeros where work happened, -0, NaN leaking into display, values exceeding a hard limit, percentages outside [0,100], totals that don't sum
  • A metric moved by an order of magnitude with no input change that explains it
  • You're about to show the output to someone who will make a decision from it

Don't use when: the output has no ground truth to compare against and no internal consistency to violate (pure creative output). Don't let it become paralysis — the skill is a screen, not a proof of correctness for every digit.

Instructions

1. Read the output as the skeptic, not the author

Before celebrating that output exists, actually read it. Ask the questions a domain reviewer would: Is this physically/financially/logically possible? Does it match the order of magnitude I'd estimate on a napkin? Is it consistent with the last known-good run? Do the parts sum to the whole?

Read the full file on GitHub · 72 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. 11d ago First seen · 72 lines · 165 tokens per session scan A 1f0f67411ec6

Subscribe to this mod's changes

smell-the-numbers is a skill published in the GitHub repository mikestangdevs/craft-skills (4 stars, last pushed 3mo ago), licensed MIT. It adds 165 tokens to every session and 1,220 once invoked, about $0.0008 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.

Related

Other skills, from other repositories

absolute-simplify

Use when the user wants to simplify, clean up, refactor, tidy, or refine code — their staged/unstaged git changes or a target file/path. Reduces complexity, flattens nesting, removes redundancy and dead code, scores each change by value (holding low-value churn), then runs tests to prove nothing broke. Invoke on…

maddhruv/absolute · 178 tokens

plumb-line-audit

Use when auditing a diff or repository against the plumb-line principles — finds laundered uncertainty, boundary leaks, hardcoded priors, overstated maturity, outputs lacking recorded lineage, and baseline drift with no explanation. Read-only: it reports, never auto-fixes.

slopstopper/plumb-line · 0 tokens

plumb-line-bootstrap

Use when setting up a project with the plumb-line discipline — interviews the builder to find their source-truth layer and layering, generates a domain-neutral ruleset, and installs parameterized enforcement (boundary check, test gate, pre-commit gate, branch guard) for the project's language. Ships no default layers…

slopstopper/plumb-line · 73 tokens

plumb-line-remediate

Use when applying findings from a plumb-line audit report — the builder has a report (or pasted findings) and wants the fixes made. Opt-in and separate from the audit, which is read-only and never fixes.

slopstopper/plumb-line · 51 tokens

plumb-line-adopt

Use when a builder wonders what plumb-line would do for their codebase or which part to adopt — or when, mid-task, their work shows a fit signal (adding a mock or fallback near a production path, mixing fixture, cached, or LLM/agent-produced data with real data) and visible uncertainty would help. Inspects the repo…

slopstopper/plumb-line · 0 tokens

plumb-line-method

Use when a builder wants to learn or be reminded of the plumb-line method — the discipline of epistemic honesty enforced by tooling. Teaches the thesis, the nine portable principles, the maturity vocabulary, and the one-line test. Pure knowledge; takes no actions.

slopstopper/plumb-line · 59 tokens