About this skill

Problem

Many users only have a partial title and still need a plain TXT file that can be saved locally. Search results often mix HTML pages, directory listings, paid platforms, redirects, and files with GBK or GB2312 encoding, so a naive download may produce an error page or unreadable text. This skill turns that cleanup loop into a repeatable workflow.

How It Works

It starts from the novel name, uses web_search to find public free sources, and prefers direct TXT links, raw.githubusercontent.com, known free novel sites, or document-library pages. Candidate URLs are checked with web_fetch to see whether they are direct files, contain .txt listings, or are only search pages. Once a usable link is found, the file is saved to /tmp/ first, then validated for size, content, and encoding. If the text is garbled, it attempts iconv conversion to UTF-8, then saves the result to a user-specified directory or the default ~/Downloads.

Boundaries

The skill is intended for public free sources only. It should not be used with paid e-book platforms, DRM, or paywall bypass, and it is not suited for mass redistribution. If a site blocks access, redirects look suspicious, or the file is under 10KB, treat it as a failed source and retry with another candidate. For multi-volume works, it downloads the most complete volume first and reports the rest.

Screenshots

Use Cases

Before offline reading, find a downloadable TXT for a named Chinese web novel and save it as UTF-8.
Given a partial title, distinguish directory pages, search pages, and direct links, then fetch the best volume.
When a downloaded novel is garbled GBK text, verify the file is over 10KB and convert it to UTF-8.
While building an English novel corpus, grab TXTs from public free sources and avoid paid, DRM, or HTML pages.

Best For

Text researchers annotating Chinese web novels offline, who need editable TXT files with clean encoding.
Data engineers preparing corpora, who need novel body text from public free sources rather than HTML pages.
Technical readers using local readers, who want Chinese and English novels saved as UTF-8 TXT files.
Automation engineers maintaining reading toolchains, who need checks for direct TXTs, listings, redirects, and bad sources.

Related Skills

Text Classifier Pro

@user_26bdbd50/text-classifier-pro

Batch-classify pasted text or CSV rows with preset and custom templates, outputting structured labels and confidence scores for data organization and annotation workflows.

Data Analysis 2026.08.30 ↓ 598

Bazhuayu Global Public Opinion and Brand Reputation Analysis

@user_2d9d1ec9/bazhuayu001

Analyzes product VOC and brand reputation events across public news, community, and social pages, delivering Markdown, Excel, and optional HTML reports.

Data Analysis 2026.08.30 ↓ 600