# AI Slop Patterns Explicitly Banned by Diagram Design: Complete Taxonomy List

> Diagram Design explicitly bans eight AI slop patterns like type-llm and type-gpt. Discover the complete taxonomy list to ensure your diagrams meet repository quality standards.

- Repository: [Cathryn Lavery/diagram-design](https://github.com/cathrynlavery/diagram-design)
- Tags: deep-dive
- Published: 2026-09-10

---

**Diagram Design prohibits eight specific schematic patterns—including `type-llm`, `type-gpt`, and `type-neural`—defining any diagram tagged with these identifiers as AI slop that violates repository quality standards.**

The `cathrynlavery/diagram-design` repository maintains strict semantic taxonomy rules by enforcing the **AI slop patterns explicitly banned by Diagram Design**, preventing generic machine-generated content from polluting its curated diagram library. According to [`skills/diagram-design/SKILL.md`](https://github.com/cathrynlavery/diagram-design/blob/main/skills/diagram-design/SKILL.md), the system automatically rejects any schematic labeled with AI-derived source tags, ensuring only intentional, human-verified architectural patterns enter the collection.

## The Eight Banned AI Slop Patterns

In [`skills/diagram-design/SKILL.md`](https://github.com/cathrynlavery/diagram-design/blob/main/skills/diagram-design/SKILL.md), the repository maintains a definitive blacklist of prohibited pattern types. These **AI slop patterns explicitly banned by Diagram Design** are automatically flagged during validation:

- **`type-llm`** — Generic Large Language Model outputs without human architectural oversight
- **`type-chatgpt`** — Diagrams produced by ChatGPT-style conversational agents
- **`type-gpt`** — Any schematic directly attributed to GPT model generation
- **`type-openai`** — Content sourced from OpenAI services or API endpoints
- **`type-anthropic`** — Diagrams generated by Anthropic models or Claude-based systems
- **`type-ml`** — General machine learning-generated schematics
- **`type-neural`** — Neural network-driven diagram outputs
- **`type-ai`** — Catch-all tag for any unspecified AI-generated content

These identifiers appear in the AI slop section of the skill definition file, where they are explicitly marked as prohibited taxonomy entries that trigger automatic rejection.

## How the Ban Is Enforced

The repository validates incoming diagrams against this banned list during the ingestion process. When a schematic file contains any of the prohibited type declarations, the system categorizes it as AI slop and rejects the submission.

```yaml

# ❌ REJECTED: Contains banned AI slop pattern

type: type-gpt
title: "Microservice Architecture from ChatGPT"
content: |
  [Generated diagram content]

```

In contrast, human-curated or approved pattern types follow the allowed taxonomy:

```yaml

# ✅ ACCEPTED: Uses approved semantic pattern

type: type-architecture
title: "Microservice Architecture"
content: |
  [Curated diagram content]

```

## Rationale: Preserving Semantic Integrity

The architectural decision record in [`docs/adr/0002-semantic-patterns-do-not-expand-the-taxonomy.md`](https://github.com/cathrynlavery/diagram-design/blob/main/docs/adr/0002-semantic-patterns-do-not-expand-the-taxonomy.md) explains why these AI slop patterns remain excluded from the taxonomy. The document establishes that Diagram Design prioritizes **intentional, human-verified architectural patterns** over automatically generated variations that lack semantic precision.

By banning `type-llm`, `type-neural`, and related identifiers, the repository prevents dilution of the curated pattern library with generic outputs that may contain hallucinated architectural components or inconsistent notation standards.

## Summary

- **Eight specific patterns** constitute AI slop in Diagram Design: `type-llm`, `type-chatgpt`, `type-gpt`, `type-openai`, `type-anthropic`, `type-ml`, `type-neural`, and `type-ai`
- The canonical blacklist resides in [`skills/diagram-design/SKILL.md`](https://github.com/cathrynlavery/diagram-design/blob/main/skills/diagram-design/SKILL.md), which defines prohibited schematic sources
- Validation logic rejects any diagram tagged with these identifiers during ingestion to preserve taxonomy integrity
- [`docs/adr/0002-semantic-patterns-do-not-expand-the-taxonomy.md`](https://github.com/cathrynlavery/diagram-design/blob/main/docs/adr/0002-semantic-patterns-do-not-expand-the-taxonomy.md) documents the architectural decision to exclude AI-generated content from pattern expansion

## Frequently Asked Questions

### What exactly qualifies as "AI slop" in Diagram Design?

AI slop refers to any schematic explicitly tagged with the eight banned pattern types listed in [`skills/diagram-design/SKILL.md`](https://github.com/cathrynlavery/diagram-design/blob/main/skills/diagram-design/SKILL.md). The repository treats these as low-quality, machine-generated outputs that bypass human curation, including diagrams sourced from ChatGPT, GPT-4, Claude, or other large language models.

### Why does Diagram Design ban AI-generated patterns from its taxonomy?

According to [`docs/adr/0002-semantic-patterns-do-not-expand-the-taxonomy.md`](https://github.com/cathrynlavery/diagram-design/blob/main/docs/adr/0002-semantic-patterns-do-not-expand-the-taxonomy.md), the ban preserves semantic precision in the architectural pattern library. AI-generated diagrams often contain hallucinated components or inconsistent notation that would dilute the quality of the curated collection.

### Can I submit a diagram if I used AI tools during creation but remove the AI tag?

The source code in [`skills/diagram-design/SKILL.md`](https://github.com/cathrynlavery/diagram-design/blob/main/skills/diagram-design/SKILL.md) indicates that validation specifically checks for the prohibited identifiers: `type-llm`, `type-gpt`, `type-chatgpt`, `type-openai`, `type-anthropic`, `type-ml`, `type-neural`, and `type-ai`. If your schematic uses approved type declarations such as `type-architecture` and meets the repository's semantic standards, it may pass validation, though the underlying ADR emphasizes preference for human-verified architectural intent.

### Where is the complete list of banned patterns documented?

The complete list appears in [`skills/diagram-design/SKILL.md`](https://github.com/cathrynlavery/diagram-design/blob/main/skills/diagram-design/SKILL.md) under the AI slop section. This file serves as the source of truth for prohibited schematic types, while [`docs/adr/0002-semantic-patterns-do-not-expand-the-taxonomy.md`](https://github.com/cathrynlavery/diagram-design/blob/main/docs/adr/0002-semantic-patterns-do-not-expand-the-taxonomy.md) provides the architectural rationale for maintaining these restrictions.