Skip to content
Back to search

unslop

Empirically detect the repetitive defaults a model falls back on, then turn those findings into a reusable instruction file that makes future outputs less generic.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/mshumer--unslop/badge.svg)](/listings/mshumer--unslop)

HTML

<a href="/listings/mshumer--unslop"><img src="/api/listings/mshumer--unslop/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

You need [Claude Code](https://docs.anthropic.com/en/docs/claude-code) installed.

git clone https://github.com/mshumer/unslop.git
cd unslop

python3 -m venv .venv
source .venv/bin/activate

# Text / writing domains
python3 unslop.py --domain "blog writing"

# Visual / website domains
pip install playwright && playwright install chromium
python3 unslop.py --domain "startup SaaS landing pages" --type visual --count 20 --concurrency 3

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 114 lines
  1. # unslop
  2. Empirically detect the repetitive defaults a model falls back on, then turn those findings into a reusable instruction file that makes future outputs less generic.
  3. ## Send This To Your Agent
  4. Copy and send this sentence:
  5. ```text
  6. Clone https://github.com/mshumer/unslop, read https://github.com/mshumer/unslop/blob/main/skills/unslop/SKILL.md, run unslop for DOMAIN_HERE, review unslop-output/analysis.md and unslop-output/skill.md for specificity, and return the finished skill file plus the main repeated patterns you found.
  7. ```
  8. Agent-readable instructions live here: [`skills/unslop/SKILL.md`](skills/unslop/SKILL.md)
  9. ## What It Does
  10. Every model has defaults it collapses toward. In writing, that means the same openings, transitions, and rhetorical moves. In visual work, that means the same landing-page structure, the same gradients, the same card grids, the same CTA blocks.
  11. `unslop` measures those defaults instead of guessing at them.
  12. 1. You give it a domain.
  13. 2. It generates a prompt set and produces many samples with Claude Code.
  14. 3. For visual work, it renders screenshots so the analysis can inspect what the pages actually look like.
  15. 4. It analyzes the whole set for repeated patterns.
  16. 5. It writes `skill.md`, a focused file of what to avoid next time.
  17. 6. It runs a before/after comparison so you can see whether the profile actually changed the result.
  18. The main output is a single markdown file you can reuse in Claude Code, `CLAUDE.md`, Codex, Cursor rules, or any system prompt.
  19. ## Quickstart
  20. You need [Claude Code](https://docs.anthropic.com/en/docs/claude-code) installed.
  21. ```bash
  22. git clone https://github.com/mshumer/unslop.git
  23. cd unslop
  24. python3 -m venv .venv
  25. source .venv/bin/activate
  26. # Text / writing domains
  27. python3 unslop.py --domain "blog writing"
  28. # Visual / website domains
  29. pip install playwright && playwright install chromium
  30. python3 unslop.py --domain "startup SaaS landing pages" --type visual --count 20 --concurrency 3
  31. ```
  32. ## Output
  33. Results land in `./unslop-output/`:
  34. ```text
  35. unslop-output/
  36. skill.md
  37. analysis.md
  38. prompts.json
  39. samples/
  40. screenshots/
  41. before-after/
  42. before.md / before.html
  43. after.md / after.html
  44. before.png / after.png
  45. ```
  46. `skill.md` is the file you usually want to keep and reuse.
  47. ## Common Commands
  48. ```bash
  49. python3 unslop.py --domain "marketing emails" --count 50 --concurrency 5
  50. python3 unslop.py --domain "startup SaaS landing pages" --type visual --count 20 --concurrency 3
  51. python3 unslop.py --domain "Python tutorials" --count 100 --concurrency 10
  52. python3 unslop.py --domain "React landing pages" --type visual --model sonnet --effort low
  53. ```
  54. ## Options
  55. | Flag | Default | Description |
  56. |---|---|---|
  57. | `--domain` | required | What to analyze |
  58. | `--type` | `text` | `text` for writing/code/prose, `visual` for websites/HTML |
  59. | `--count` | `50` | Number of samples to generate |
  60. | `--concurrency` | `5` | Parallel Claude Code calls |
  61. | `--model` | Claude default | Claude model alias/name to pass through |
  62. | `--effort` | Claude default | Claude effort level to pass through |
  63. | `--timeout` | `600` | Seconds to wait per Claude call |
  64. | `--analysis-timeout` | `1800` | Seconds to wait for the analysis pass |
  65. | `--retries` | `1` | Retries per failed Claude call |
  66. | `--output` | `./unslop-output` | Where to put results |
  67. | `--skip-comparison` | `false` | Skip the before/after step |
  68. ## Prebuilt Profiles
  69. The `profiles/` directory contains ready-to-use profiles:
  70. - [`profiles/writing.md`](profiles/writing.md)
  71. - [`profiles/react-design.md`](profiles/react-design.md)
  72. Run the tool yourself when you want something narrower or domain-specific.
  73. ## Philosophy
  74. `unslop` focuses on what to avoid, not on prescribing a single new template. The point is to remove the defaults that keep repeating and force the model to make fresher choices.
  75. ## Requirements
  76. - [Claude Code](https://docs.anthropic.com/en/docs/claude-code)
  77. - Python 3.10+
  78. - For visual domains: `pip install playwright && playwright install chromium`
  79. ## License
  80. MIT

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.