X Post Reader — Communication skill for Claude Code

x-post-reader

A drop-in agent skill for **reading X/Twitter posts** without an API key, login, or paid plan — plus optional keyword search when you have a key.

一个给 Claude Code / Codex / DSH 等 agent 用的技能：**无 key、无登录读取 X/Twitter 推文**（含长文 X Article），可选关键词搜索。

Why

Fetching `x.com` directly fails: the page is a JavaScript shell that needs login, and a plain HTTP client gets nothing (or, in mainland China, the host is unreachable). This skill routes around that with public JSON mirror APIs that return the full post — text, author, engagement, media, and the complete body of long-form posts (X Articles).

What it can do

Capability	Free	How
Read a post by URL or id	✅	api.fxtwitter.com//status/
Full text of long-form posts	✅	included in the fxtwitter JSON
User profile	✅	api.fxtwitter.com/
Batch-hydrate many URLs	✅	bundled script
Keyword search	🔑	needs an X API v2 bearer (or third-party)
Replies / comments	❌	not available without login

Install

Copy the repo into your agent's skills directory:

# Claude Code
cp -r x-post-reader ~/.claude/skills/

# Codex
cp -r x-post-reader ~/.agents/skills/

# DeepSeek Harness (dsh)
cp -r x-post-reader ~/.agents/skills/

Or install via your agent's skill-install command. The `SKILL.md` carries the `name`/`description` frontmatter every agent reads when deciding what to load.

Usage

python scripts/xpost.py tweet https://x.com/yaohui12138/status/2097151445883899913
python scripts/xpost.py tweet 2104912748794589515          # bare id works too
python scripts/xpost.py user xilo2991
python scripts/xpost.py batch urls.txt
python scripts/xpost.py search "AI agent" --bearer $X_BEARER_TOKEN

No dependencies — Python 3 standard library only. `HTTP_PROXY` / `HTTPS_PROXY` (and the Windows system proxy) are honoured automatically.

How it works

1. Parse the tweet handle + id out of any x.com / twitter.com URL (o