A rigorous quality assurance framework designed to detect and eliminate low-value AI-generated content patterns while ensuring factual accuracy.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install antislop
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install antislop using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Anti-Slop?
Anti-Slop is a specialized skill designed to combat the phenomenon of slop—AI-generated content that appears productive but lacks genuine value, insight, or information density. By applying this skill within your Openclaw Skills workflow, you can filter out generic statements, filler phrases, and repetitive structures that plague standard LLM outputs. It forces a higher standard of original work by prioritizing specific evidence and domain-specific knowledge over surface-level analysis.
This skill acts as a technical gatekeeper, ensuring that every sentence produced serves a specific purpose. It focuses on identifying and correcting common failure modes such as confident hallucinations, unverified success claims, and excessive formatting. For developers and researchers using Openclaw Skills, this represents a shift from quantity to quality, moving toward more reliable and actionable automation.
Anti-Slop Use Cases
- Reviewing automated pull request descriptions to ensure they contain specific code references instead of generic summaries.
- Auditing technical documentation generated via Openclaw Skills for information density and original insight.
- Verifying the success of automated tasks by cross-referencing claims with actual CLI tool outputs.
- Cleaning up corporate filler and redundant phrases from generated reports or communications.
- Ensuring AI-generated analysis of codebases is based on actual file inspection rather than memory patterns.
How Anti-Slop Works
- The Information Density Test evaluates every sentence to determine if it adds unique value or serves as filler.
- The Originality Test distinguishes between genuine insights derived from investigation and generic advice found in training data.
- The False Success Audit verifies every claim of task completion by demanding specific evidence like file paths or test logs.
- The Structure Audit simplifies the output by removing unnecessary tables, bullet points, and headers that obscure the core message.
- The Plagiarism Check ensures the output is not a reworded template or a direct copy of common documentation patterns.
Anti-Slop Setup
To integrate this quality control protocol into your local environment, ensure your agent has access to the necessary verification tools.
# Enable the Anti-Slop reviewer within your project
openclaw skill install anti-slop
# Configure severity levels for automated gatekeeping
openclaw config set anti_slop.min_severity medium
# Run a manual quality check on a generated file
openclaw run anti-slop --file path/to/output.md
Anti-Slop Data Schema & Taxonomy
The skill organizes its quality assessments into a structured taxonomy to help users prioritize improvements:
| Level | Criteria | Required Action |
|---|---|---|
| Critical | Hallucinations, false claims, or phantom references. | Immediate fix required before output. |
| High | Unverified success claims or stale technical data. | Verify using live tools or remove. |
| Medium | Low density, filler text, or meta-commentary. | Rewrite to increase density. |
| Low | Unnecessary formatting or minor stylistic issues. | Refine if time permits. |
Anti-Slop Advanced Features
- Real-time verification hooks that trigger tool execution (like gh pr view) to validate agent claims.
- Multi-stage auditing that classifies issues by severity level for automated CI/CD gating.
- Custom pattern matching to detect and eliminate project-specific jargon or recurring AI clichés.
- Density-scoring algorithms that provide a quantitative measure of information per paragraph.
- Integration with Openclaw Skills to provide automated feedback loops for self-correcting agents.
SKILL.md
Loading
Related Openclaw Skills
Fact Check
A systematic verification tool for AI coding agents to eliminate hallucinations by validating technical claims against real-world data.
Foreign Trade News Aggregator
An automated intelligence tool that aggregates foreign trade news via RSS, translates headlines into Chinese, and delivers reports to Feishu.
TikTok Comment Scraper
A high-performance scraping tool designed to extract comments and threaded replies from TikTok videos and photo posts without requiring a login.
Content Writer Pro
A professional AI-driven copywriting generator designed to create high-converting marketing material, social posts, and brand stories.
Git Daily Report & Code Reviewer
An automated agent skill that generates daily Git activity reports and performs security-focused code reviews across multiple repositories.
PostQube Threads Publisher
A comprehensive social media automation skill for publishing posts and native thread storms to Threads via the PostQube API.
Featured*
