- Home
- Skills
- Technical Documentation
- DESLOP-THIS
DESLOP-THIS
Strip AI tells and robotic patterns from text using a dual-layer linter and stylistic judgment process.
Free
Works with the AI Agents you already use
DESLOP-THIS
Example session with this skill installed
Run the deslop skill on this cover letter draft. It feels too generic and robotic. Use the email register.
- Read your context and instructions
- Compiled the deslop-this
I processed the draft using the email register. High-tier tells like "tapestry" and "delve" were removed, and I've adjusted the sentence rhythm to break the 18-word average.
Revised Cover Letter
I'm writing to apply for the Lead DevRel role. I've followed your recent work on the V3 migration, specifically the decision to drop the legacy auth layer. It's a move most teams are too scared to make.
I spent three years at Orbit building their docs from zero. I didn't just write guides; I built the telemetry system that tracked how users actually moved through the onboarding flow. We found that 40% of users dropped off at the CLI setup, so I rewrote the tool's init logic. It wasn't about "navigating a landscape." It was about fixing a broken shell script.
I don't believe in passive community management. I believe in shipping code that makes the community's life easier. I'd like to show you the internal dashboard I built for tracking contributor sentiment.
When can we talk?
Next steps
- Run
node .claude/skills/deslop/lint.mjson this version to verify the final rhythm stats. - Replace the "Orbit" and "40%" placeholders with your actual previous company and metric if different.
- Check the whitelist in
ai-tells.jsonif "telemetry" is flagged as a domain-specific term you want to keep.
Connects securely to your tools. The creator never sees your data.
About this skill
AI writing often suffers from recognizable patterns: mid-length sentences, vague metaphors, and repetitive structures. This skill strips away these "tells" by combining a deterministic linter with human-like judgment rules. It forces text out of the generic AI band and into a natural, specific register.
What it does
- Automated Linting identifies high-risk "AI words" and phrases across three severity tiers.
- Register Enforcement applies specific constraints for job applications, emails, or long-form prose.
- Rhythm Analysis calculates sentence length statistics to break the 15-20 word "AI band" monotony.
- Structural Correction targets non-regex patterns like both-sidesing, tidy anecdotes, and anaphora.
- Texture Injection inserts specific markers like contractions and named entities to ground the writing.
How it works
- Run
lint.mjsto get a report of HIGH, MED, and LOW findings with specific line-item fixes. - Apply Judgment Rules to catch conceptual tells like invented labels or dead metaphors.
- Perform the Addition Pass to add texture via real numbers, opinions, or named tools.
- Fix Rhythm by varying sentence lengths to meet standard deviation targets.
Frameworks & tools
The skill relies on a standalone Node.js linter (lint.mjs) compatible with Node 18+, a JSON-based word tier library (ai-tells.json), and a reference guide for stylistic judgment.
Why this beats prompting it yourself
Standard "humanize" prompts often just swap common words for obscure ones, which looks worse. This tool uses a mechanical linter to catch objective tells and a structured four-pass process to ensure the rhythm and specific details are actually present.
Use cases
- Application answers for grant forms or job intakes where robotic phrasing is a hard fail.
- Cold outreach emails that need to avoid the typical "I hope this finds you well" structure.
- Technical READMEs that need to sound like they were written by the developer who built the tool.
Known limitations
Requires Node.js 18+ to run the linter. Does not support scattering typos as a humanization tactic. Facts must be provided by the user as the skill will not invent data to satisfy texture markers.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Technical Documentation.
Requires Node.js 18+. Will not invent facts for texture markers. Register constraints are strict (e.g., no emoji in emails/answers). Exit code 1 on HIGH severity findings.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
1 use
Times used to date
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 24 days ago
- Free with an Agensi account
- Software
Git