x-read — Claude Code command: analyze X posts and threads
x-read — a command that performs deep analysis of X (Twitter) posts or threads using Grok. Extracts the original text, thread chain, key claims, reply sentiment, and notable voices in the conversation.
MIT ★ 5,412 updated 2026-10-24
/plugin marketplace add eugeniughelbur/obsidian-second-brain
/plugin install obsidian-second-brain@obsidian-second-brainWhat x-read does
The /x-read [url] command accepts a link to an X post and runs the scripts.research.x_read script from the skill root. It returns a structured analysis: original post, full thread, summary, extracted claims, reply sentiment patterns, notable counterarguments, and voices worth following.
The analysis displays in chat as-is without summarization. By default, the output stays in the conversation only. To archive it in your vault, request explicitly (e.g., "save this to vault"). The command works with any URL containing x.com/ or twitter.com/.
For Claude Code, installation requires two steps: run /plugin marketplace add eugeniughelbur/obsidian-second-brain, then /plugin install obsidian-second-brain@obsidian-second-brain. Also include the .claude/commands/x-read.md file from the repository.
What it helps with
- Read a complete thread without scrolling through X manually
- Surface key claims and the counterarguments they receive
- Identify authoritative voices participating in the discussion
- Assess the overall sentiment of replies to a post
- Archive analysis for future reference
Example prompts for the agent
- Deep read this X post for me
- Analyze this thread
- What are people saying about this post
How to install x-read
1Plugin marketplace
/plugin marketplace add eugeniughelbur/obsidian-second-brain/plugin install obsidian-second-brain@obsidian-second-brain2Manually: put the files in place
.claude/commands/x-read.mdCall it with /x-read. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/x-read.mdCall it with /prompts:x-read. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/x-read.prompt.mdManually: put the files in place
.gemini/commands/x-read.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about x-read
How do I install x-read in Claude Code?
Run /plugin marketplace add eugeniughelbur/obsidian-second-brain, then /plugin install obsidian-second-brain@obsidian-second-brain. Copy the .claude/commands/x-read.md file from the repository into your project's .claude/commands/ directory.
What does x-read do?
It analyzes an X post or thread using Grok x_search, showing the original text, complete reply chain, key claims, reply sentiment, and important voices in the discussion.
Is the analysis saved automatically?
No. By default, results appear only in chat. To save the analysis to your vault, ask explicitly — for example, "save this to vault".
Source text
x-read.md · MIT
---
description: Deep-read an X (Twitter) post via Grok x_search - verbatim post, thread, TL;DR, claims, reply sentiment, voices to watch
category: research
triggers_en: ["read this x post", "deep read this tweet", "analyze this tweet", "read this thread"]
triggers_de: ["diesen X-Beitrag lesen", "diesen Tweet gründlich lesen", "diesen Tweet analysieren", "diesen Thread lesen"]
triggers_es: ["léeme este post de x", "profundiza en este tweet", "analiza este tweet", "léeme este hilo"]
triggers_pt: ["leia este post do x", "leitura profunda deste tweet", "analise este tweet", "leia esta thread"]
triggers_zh: ["读一下这条 X 帖子", "深入分析这条推文", "这条推文讲了什么", "梳理一下这串推文"]
---
Use the obsidian-second-brain skill. Execute `/x-read [url]`:
1. Resolve the URL from the user's argument. If no URL was given, ask: "Which X post URL?" Accept any URL containing `x.com/` or `twitter.com/`.
2. Run the script from the skill root (its absolute path was given at session start as **Skill root**; substitute it for `SKILL_ROOT`):
```bash
uv run --directory "SKILL_ROOT" -m scripts.research.x_read "<url>"
```
3. The script prints a structured analysis (ORIGINAL POST, THREAD, TL;DR, KEY CLAIMS, REPLY SENTIMENT, NOTABLE COUNTER-ARGUMENTS, VOICES TO WATCH) and a one-line cost summary on stderr. Show the analysis to the user verbatim - don't paraphrase or summarize.
4. **Default save behavior: chat only.** Do NOT save the analysis to the vault automatically. The user must ask explicitly ("save this", "save to vault", "/obsidian-save") for it to be archived.
5. If the user asks to save: write an AI-first note to `Research/X-reads/YYYY-MM-DD - <slug>.md` in the vault, following the AI-first vault rule (Section 0 of `_CLAUDE.md`):
- Frontmatter: `date`, `time`, `type: x-read`, `ai-first: true`, `post-url`, `post-author` (if known), `key-claims` (list), `tags`, `related-people` (wikilinks for any @ handles that map to known people in the vault), `cost-usd`
- Body starts with a `## For future agent` preamble section (2-3 sentences summarizing what this post is about and why it was saved)
- Then the full structured analysis from the script
6. Plain English triggers that route to this command: "read this tweet", "read this X post", "what's in this tweet", "analyze this X link" - when followed by a URL.
7. If the script fails with a clear error (missing key, network down), surface the error message verbatim. The script handles retry on transient errors automatically.
---
**AI-first rule:** Every note created or updated by this command MUST follow `references/ai-first-rules.md` - `## For future agent` preamble, rich frontmatter (`type`, `date`, `tags`, `ai-first: true`, plus type-specific fields), recency markers per external claim, mandatory `[[wikilinks]]` for every person/project/concept referenced, sources preserved verbatim with URLs inline, and confidence levels where applicable. If that path does not resolve from your working directory, search upward for it; if you still cannot read it, say so before writing rather than producing a note that silently skips the rule. The vault is for future agent retrieval - not human reading.
**Anti-fabrication:** Search exhaustively before claiming any note, person, or file is absent - false absence is the most common failure mode - and never invent facts, entities, or dates (mark unknowns as `TBD`). See the anti-fabrication and search-completeness hard rules in `references/ai-first-rules.md`.
© Eugeniu Ghelbur · MIT · eugeniughelbur/obsidian-second-brain