slop-cop · workspace mirror

Audit your prose for AI slop & reader comprehension.

A read-only mirror of the slop-cop rule set. @-mention this project from any other Lovable project. The agent loads the rules, runs the dual audit, returns a verdict.

45
AI-slop patterns
See the file →
150+
Vocabulary tells
See the file →
35
Comprehension tells
See the file →
10
Rule files mirrored
See the file →

Quick start

three steps, two minutes
  1. 01

    Put both projects in the same Lovable workspace.

  2. 02

    In the consuming project's chat, type @ and pick this project.

  3. 03

    Ask the agent to audit. It reads SKILL.md + the relevant references and returns a verdict.

The rule set

All rules →
MetaSKILL.md
SKILL.md

Entry point for the agent. Modes (audit vs fix), when to skip, the workflow, and the top-20 lethal items.

Read →
MetaCalibration
references/calibration.md

Density formula, severity tiers, genre adjustments, model fingerprints, contested-tell handling. Read this first.

Read →
AI-SlopPatterns
references/patterns.md

45 rhetorical and structural patterns across 5 groups. The bulk of the AI-slop axis.

Read →
AI-SlopVocabulary
references/vocabulary.md

~150 word-level tells across 7 categories: LLM verbs, cliché metaphors, intensifiers, sycophancy, weasels, connectors, spike words.

Read →
AI-SlopFormatting tells
references/formatting-tells.md

~33 markdown and structural patterns: bold-first bullets, title clichés, repetition, whitespace artifacts.

Read →
ComprehensionComprehension
references/comprehension.md

35 patterns across 5 groups for the comprehension axis: density overload, telegraphic compression, audience assumptions, scannability, cognitive friction.

Read →
ComprehensionReadability metrics
references/readability-metrics.md

Eight readability formulas with audience targets and the three cold-reader density signals (acronyms, named entities, numerics).

Read →
MetaAudit report template
references/audit-report-template.md

Canonical dual-verdict output: AI-slop and Comprehension scores, top fixes, mechanical and qualitative violations, recommended action.

Read →
MetaSources
references/sources.md

~50 published sources backing every pattern. Peer-reviewed linguistics, vendor methodology, practitioner literature, viral takedowns.

Read →
Scannerscan.py
scripts/scan.py

Python regex scanner. Density-aware scoring, genre flags, model fingerprinting. JSON / quick / human modes.

Read →
*

Density is the signal.

Single instances of delve or one em dash in 1,000 words don't trip the verdict — concentration does. The pinned upstream commit lives in LINEAGE.md.