CtrlK
BlogDocsLog inGet started
Tessl Logo

defuddle

Extract clean markdown content from web pages using Defuddle CLI, removing clutter and navigation to save tokens. Use instead of WebFetch when the user provides a URL to read or analyze, for online documentation, articles, blog posts, or any standard web page.

Install with Tessl CLI

npx tessl i github:kepano/obsidian-skills --skill defuddle
What are skills?

100

Does it follow best practices?

Validation for skill structure

SKILL.md
Review
Evals

Evaluation results

100%

65%

Developer Documentation Research Summary

Defuddle CLI preference over WebFetch

Criteria
Without context
With context

Uses defuddle CLI

0%

100%

Markdown flag on every call

0%

100%

No HTML output files

100%

100%

Global npm install

0%

100%

Uses -o flag to save content

0%

100%

Three page files produced

100%

100%

Summary report produced

100%

100%

fetch_content.sh present

100%

100%

Install before fetch

0%

100%

Does not use WebFetch or HTTP client

0%

100%

Without context: $1.1160 · 5m 29s · 34 turns · 280 in / 10,483 out tokens

With context: $0.5567 · 2m 1s · 33 turns · 192 in / 5,780 out tokens

55%

35%

Resource Catalog Builder

Metadata extraction with -p flag

Criteria
Without context
With context

Uses -p title

0%

0%

Uses -p description

0%

0%

Uses -p domain

0%

0%

Uses defuddle CLI

0%

100%

Markdown flag on full content fetches

0%

100%

npm install step

0%

100%

Catalog file produced

100%

100%

Script file present

100%

100%

Without context: $0.5050 · 2m 10s · 22 turns · 25 in / 9,497 out tokens

With context: $0.7333 · 3m 6s · 39 turns · 37 in / 9,717 out tokens

85%

60%

Content Archiving Pipeline

Output format selection (--md and --json)

Criteria
Without context
With context

Uses --md for markdown output

0%

100%

Uses --json for structured output

0%

100%

No bare HTML output

100%

100%

Uses defuddle CLI

0%

33%

Markdown files are clean

0%

100%

JSON files are valid JSON

100%

50%

npm install step

0%

100%

Files saved with -o flag

0%

100%

Without context: $0.3382 · 1m 56s · 19 turns · 74 in / 5,437 out tokens

With context: $0.5408 · 2m 31s · 33 turns · 282 in / 6,044 out tokens

Evaluated
Agent
Claude Code

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.