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 numen-tech/slopornot --skill slop-checkgit clone --depth 1 https://github.com/numen-tech/slopornotWrote 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/numen-tech/slopornot/slop-check)<a href="https://agentmods.dev/skills/numen-tech/slopornot/slop-check"><img src="https://agentmods.dev/badge/skills/numen-tech/slopornot/slop-check/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/numen-tech/slopornot/slop-check"><img src="https://agentmods.dev/badge/skills/numen-tech/slopornot/slop-check.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 | $0.00147 | $0.04084 |
| Opus 5 | $0.00073 | $0.02042 |
| Sonnet 5 | $0.00029 | $0.00817 |
| Haiku 4.5 | $0.00015 | $0.00408 |
Grade A, and why
slop-check 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 10d 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 — 365 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Slop Check
One-shot access to Slop or Not Pro's on-device analysis. Detect AI text or images, clean AI artifacts, or score readability, then report a clear verdict. No interview, no loop.
Slash command: /slop-check [paste text | file path | image path]
Critical rules
- For
image-detectand explicitimage-score, MUST prefer the app-bundle CLI when the current client can run local shell commands. Use MCP for image operations only when CLI execution is unsupported or fails. For text, readability, cleanup, and status, try MCP before CLI. - MUST verify Pro with a real Pro-gated call. NEVER infer Pro from
slop_statusorslop status; both succeed for non-Pro users. - NEVER block on a structured-question tool. Resolve ambiguity with the Step 1 precedence rule and state the assumption, or emit the single plain-text input request. Plain text renders on every harness.
- NEVER create symlinks, edit shell rc files, or otherwise modify PATH.
Prefer the absolute app-bundle binary at
/Applications/Slop Or Not.app/Contents/MacOS/slopwhenever using the CLI. Detection, readability, cleanup, score, and status are output-only: run them immediately, no permission needed. - Detection scores are 0-1 decimals; MUST multiply them by 100 for
display. Readability values are grade/ease numbers and MUST NOT be
converted to percentages. NEVER guess a flag or JSON field path; look it
up in
references/slop-tools.md. - Image checks MUST default to
detect_image. Usescore_imageonly when the user explicitly asks for a raw or absolute OmniAID score. A request to "score this image" without OmniAID still maps toimage-detect.
Step 1: Identify the operation and input
Pick one operation from the request. Do not ask a question to disambiguate.
| Operation | Trigger language |
|---|---|
text-detect |
"is this AI", "did a bot write this", "AI written" |
image-detect |
"is this image AI", "AI generated picture", "real photo" |
image-score |
"raw OmniAID score", "absolute OmniAID model score" |
readability |
"reading level", "what grade", "Flesch", "how readable", or a native formula name ("LIX", "Wiener Sachtextformel", "Gulpease", "Flesch-Szigriszt", "Reading Ease") |
cleanup |
"clean this", "strip invisible/zero-width", "remove homoglyphs" |
status |
"is slop set up", "is Slop or Not working", "check Pro" |
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 10d ago First seen · 365 lines · 147 tokens per session scan A 818ad8517e9f
slop-check is a skill published in the GitHub repository numen-tech/slopornot (47 stars, last pushed 2mo ago), licensed MIT. It adds 147 tokens to every session and 4,084 once invoked, about $0.0007 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 2026-08-30.
Other skills, from other repositories
untell
Audit an AI detector and measure what its verdict is worth, using a closed-loop detector-feedback rewrite as the instrument (the "untell" skill). Use when the user wants to check whether text reads as AI-generated, find out why a detector flagged something, measure a detector's false-positive rate on human writing…
ai-pattern-detector
Use when the user wants writing de-AI'd or checked for AI patterns — whether they paste the text, point to a file or URL, or ask you to review a draft — to make it sound more human and less like ChatGPT-default prose. Flags telltale vocabulary, constructions, punctuation, hedges, closers, formatting, and voice tells…
humanize-korean
A Korean-language editing skill that makes AI-written text sound more natural while preserving its meaning, facts, numbers, names, quotations, and genre. It focuses on patterns such as translation-like wording, repetitive structure, and overly mechanical rhythm.
humanize-redo
A second editing pass for Korean text produced by the humanize-korean workflow. It can revisit remaining issues in selected categories or paragraphs, with adjustable editing strictness.
humanize
A command for making AI-written Korean text sound more natural through an automated editing process.
patina
Detect and rewrite AI writing patterns in Korean, English, Chinese, and Japanese text so it reads as if a human wrote it. Meaning-preservation (MPS) verified.