- Home
- Skills
- Code Quality & Review
- AI Tells Check
AI Tells Check
A pre-flight linter that flags common AI-generated "tells" and slop patterns in frontend and marketing code.
Free
Works with the AI Agents you already use
AI Tells Check
Example session with this skill installed
Run an anti-slop check on demo-page.html to see if it looks too generic.
- Read your context and instructions
- Compiled the ai tells check
I ran the pre-flight linter on demo-page.html. It found 3 AI tells, so the file fails the check with exit code 1.
AI-tells found
- 1 em-dash(es) — taste-skill bans em-dashes in UI copy; use a hyphen or restructure
- version/BETA label in hero (e.g. 'V0.6', 'BETA')
- 'quietly in use at' filler header
What this means: the page carries three of the highest-signal fingerprints of unedited AI output. The linter is read-only, so nothing was changed. Fix the three flags, then re-run check.py. A clean file prints "clean — no AI-tells" and exits 0, which is what you want before you ship.
Next steps
- Replace the em-dash with a hyphen or restructure the line.
- Remove the V0.6 BETA hero label, or use a specific version number.
- Replace the "quietly in use at" filler with a specific proof point, like a real customer count or logo.
- Re-run python3 check.py demo-page.html to confirm a clean exit code 0.
Connects securely to your tools. The creator never sees your data.
About this skill
Generic AI generation often leaves behind distinct fingerprints that scream "bot-made" to savvy users. This skill acts as a pre-flight linter that scans your frontend code and marketing copy for high-signal AI-slop patterns before you ship.
What it does
- Flags em-dashes — Detects prohibited em-dashes in UI copy that often signal unedited LLM output.
- Identifies numbered eyebrows — Catches the classic AI signature of section numbering like
06 · Introduction. - Detects placeholder labels — Spots generic
V0.xorBETAlabels frequently inserted into hero sections by generators. - Scans for filler text — Flags common phrases like
Quietly in use atand unnecessary decorative status dots. - CI-compatible auditing — Provides exit codes for automated build pipelines to prevent slop from reaching production.
How it works
- Target selection — Point the skill at a specific
.htmlor.tsxfile, or pipe content viastdin. - Pattern matching — The skill runs a non-destructive scan based on the taste-skill §9 forbidden patterns.
- Audit report — It prints a read-only list of violations.
- Decision making — The user reviews the flags to decide which generic elements to manually refine.
Frameworks & tools
This is a dependency-free Python script designed to run in any environment with Python 3. It works directly with HTML and TypeScript React (.tsx) files.
Why this beats prompting it yourself
Prompting an AI to "not look like an AI" is famously unreliable because the model often uses the very patterns you are trying to avoid. This skill uses hard-coded pattern matching to ensure these specific visual and textual "tells" never slip through unnoticed.
Use cases
- Auditing landing pages generated by v0 or similar tools for generic footprints.
- Running a final check on marketing site PRs to ensure consistent "human" tone.
- Integrating into CI/CD pipelines to block deployments containing AI-generated filler copy.
Known limitations
This is a read-only linter and will not automatically fix the code. It focuses only on specific high-signal patterns defined in the taste-skill §9 standard.
Read-only linter; does not perform auto-edits. Limited to patterns defined in taste-skill §9 (em-dashes, numbered eyebrows, specific filler phrases).
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 any payment.
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
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed today
- Free with an Agensi account