domains / documents / remove-ai-writing-patterns / unslop

unslop

Skill card

combined

The two copies describe different procedures for the same goal. The first is a manual editing skill that instructs the agent to scan, rewrite, add voice, and self-audit against a long list of AI tells. The second is a CLI-based skill that installs and pipes text through the `unslop` command for deterministic post-processing. They overlap in intent but each contains material steps the other lacks (manual pattern list vs CLI pipeline/CI integration), so crowning one would hide a real alternative.

Repositories

Other copies

Combined skill

# Unslop

Remove AI writing patterns from text before publishing, committing, or sending.

## Option A: Manual Editing (Agent-Driven)

1. Scan the text for AI patterns (puffery, AI vocabulary, em dashes, chatbot phrases, etc.).
2. Rewrite to preserve meaning while matching the intended tone.
3. Add human voice: have opinions, vary rhythm, acknowledge complexity, use first person where appropriate, and be specific.
4. Self-audit: ask "What makes this obviously AI generated?" and fix remaining tells.

## Option B: CLI Post-Processing (Deterministic)

1. Install the tool: `pipx install unslop` or `uv tool install unslop`.
2. Pipe text through the CLI: `cat draft.md | unslop --stdin --deterministic > clean.md`.
3. Use `--deterministic` for reproducible output in CI or commit hooks.
4. Review the output, as automated cleanup may occasionally alter meaning.

## Conflicts

- **michaelshimeles/skills** prescribes a manual, agent-driven editing process with a detailed pattern list.
- **sickn33/agentic-awesome-skills** prescribes a CLI-based, deterministic post-processing pipeline.

Choose Option A for nuanced, voice-driven edits. Choose Option B for automated, reproducible cleanup in pipelines.