x-reader v0.4.0 — one-sentence agent install and doctor self-check

Added

• x-reader doctor [--json]: read-only self-check that lists core and optional capabilities as OK / MISSING / BLOCKED / INFO, says who fixes each item (agent or user) and the next command, never prints secret values, and exits non-zero only when a core item is not ready. Also runnable as python3 -m x_reader.doctor.
• install.md: step-by-step setup written for AI agents, so a user can paste one sentence into Claude Code, Codex or Cursor and have the agent install, self-check and complete a first read.
• README (EN / 简体中文): one-step "paste this into your agent" setup, plus links to the product site and the hosted web reader.

Changed

• The Agent Skill bootstrap pins a verified commit that includes doctor, and names its cache venv after the pinned commit, so a new pin installs a fresh environment instead of reusing an older CLI.
• doctor repair hints reinstall the running release tag instead of moving main.

Fixed

• x-reader doctor runs on an install with missing dependencies; other commands explain what is missing and exit non-zero.
• README: WeChat uses a headless browser fallback for public articles (saved login sessions are not used yet); Xiaoyuzhou / Apple Podcasts links are read as generic web pages; the transcription model is whisper-large-v3.
• x_reader.__version__ matches the package version.