unslop-code

Detects and removes AI-generated code tells across multiple programming languages.

Updated Nov 29, 2025
One-click install
npx skills add https://github.com/achyutkneupane/Blog-Kit --skill unslop-code-achyutkneupane
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: unslop-code
Source: https://github.com/achyutkneupane/Blog-Kit/tree/main/.ai/skills/unslop-code
Command: npx skills add https://github.com/achyutkneupane/Blog-Kit --skill unslop-code-achyutkneupane

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? AI-generated code carries recognizable tells—leftover chat artifacts, placeholder comments, swallowed errors, emoji, generic names like process_data, and structural issues like boilerplate and hallucinated APIs—that make it read as machine-written and sometimes ship real bugs. This Skill identifies those tells and guides their removal so code fits the surrounding project. ## Core Features & Use Cases - Multi-language scanner: A standalone Python script scans Python, JS/TS, Java, Go, Rust, Ruby, PHP, C/C++, C#, and more, reporting each finding with file, line, severity, bug/cosmetic class, verified data share, and a fix, plus a slop score and CI-friendly exit code. - Build and audit modes: Build mode establishes the surrounding code, real requirements, and verified APIs before generation; audit mode runs build/type-check first, then the scanner, then a human diff read for structural tells. - Evidence-ranked catalog: References document 19 tells ranked by verified share from an analysis of 11,906 Reddit posts and 11,306 comments, including which tells a regex cannot catch and which popular complaints failed verification. - Use Case: Before merging an AI-assisted pull request, run the scanner with --severity high in CI to fail on leftover chat artifacts and placeholder stubs, then manually review the diff for tutorial-shaped boilerplate and hallucinated API calls. ## Quick Start Ask the AI to audit a source directory for AI-written-code tells using the unslop-code scanner and report the highest-priority fixes.

Frequently Asked Questions about unslop-code

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I scan code for AI-generated tells?▼

Run python3 scripts/unslop_code_scan.py followed by the path to a file or directory. The scanner prints each finding with file, line, severity, bug or cosmetic class, and a suggested fix, plus an overall slop score.

How do I use the scanner in a CI pipeline?▼

Run the scanner with --severity high or --json in your CI step. The exit code equals the number of high-severity findings, so the job fails automatically when conclusive AI tells like leftover chat artifacts or placeholder stubs are present.

What programming languages does the scanner support?▼

It scans Python, JavaScript, TypeScript, Java, Go, Rust, Ruby, PHP, C, C++, C#, Kotlin, Swift, and more. Comment-based tells are caught universally, while syntax-keyed rules like swallowed errors cover Python, JS/TS, Java, C#, Ruby, and Go.

Can a scanner detect hallucinated APIs or boilerplate code?▼

No. Hallucinated APIs, tutorial-shaped boilerplate, over-engineering, and repo mismatch are structural tells a regex cannot see. The skill directs you to build, type-check, and run the code first, then read the diff manually against references/tells.md.

How do I suppress a false positive from the scanner?▼

Add the text unslop-ignore anywhere on the flagged line. This marks intentional choices, such as a broad catch at a boundary or an emoji in a CLI banner, so the audit stays trustworthy.

What dependencies does the unslop-code scanner require?▼

None beyond the Python 3 standard library. The scanner is a single standalone script using only os, re, sys, json, and argparse, so it runs anywhere Python 3 is installed without pip installs.