x-to-slack

name	x-to-slack
description	Fetches tweet content from X/Twitter via FxTwitter API, performs web research, and posts structured intelligence to Slack in a 3-message thread. Use when user shares an X/Twitter link (x.com or twitter.com URL) and wants to analyze it for Slack posting, or says "tweet to slack", "share this tweet", "x-to-slack". Do NOT use for general Slack messaging, channel management, or non-Twitter content sharing. Korean triggers: "분석", "검색", "리서치", "API".
metadata	{"author":"thaki","version":"1.2.0","category":"execution"}

X-to-Slack: Tweet Intelligence Pipeline

Note: This is the project-level Twitter-only handler (v1.1.0). The canonical universal handler supporting Twitter, GitHub, YouTube (with Defuddle transcript), and articles is at ~/.cursor/skills/pipeline/x-to-slack/SKILL.md (v2.2.0). Use the user-level version for non-Twitter URLs.

Process an X (Twitter) URL, gather context through web research, and post a structured 3-message thread to a specified Slack channel.

Input

The user provides:

1. X URL -- e.g. https://x.com/user/status/1234567890
2. Slack channel name -- e.g. #general or general

Workflow

Step 0.5: Cross-Repo Dedup Check (MANDATORY)

Before any processing, check the central intelligence registry to avoid duplicate Slack posts across repos/machines.

python3 "scripts/intelligence/intel_registry.py" check "<tweet_url>"

• Exit 0 (new): Proceed with the pipeline.
• Exit 1 (duplicate): Report "이미 처리된 URL입니다 (처음 처리: {first_seen}, 소스: {source_repo})" and STOP. Do NOT post to Slack.

If intel_registry.py is not found, log a warning and proceed (graceful degradation).

Step 1: Fetch Tweet via FxTwitter API

Convert the URL by replacing x.com or twitter.com with api.fxtwitter.com. Use WebFetch to retrieve the JSON.

For response schema, field extraction details, and error codes, see references/fxtwitter-api.md.

Extract: tweet.text, tweet.author.name, tweet.author.screen_name, tweet.likes, tweet.retweets, tweet.views, tweet.url, tweet.quote, tweet.media.

If code is not 200, inform the user of the error and do not proceed.

Step 2: Additional Web Research

Based on the tweet content:

1. Identify 2-3 key topics, technologies, or entities mentioned.
2. Run WebSearch for each to gather background, recent developments, and implications.
3. Collect relevant findings for the summary.

Step 3: Find Slack Channel

Step 3a: Channel Registry Lookup (fast path)

Strip the # prefix from the user-provided channel name and look it up in the registry below. If found, use the channel_id directly — skip the MCP search entirely.

Channel Name	Channel ID	Type	Description
research	C0A7GBRK2SW	private	리서치/논문/기술 분석
ai-coding-radar	C0A7K3TBPK7	public	AI 코딩 에이전트/도구 동향
press	C0A7NCP33LG	public	언론/미디어/콘텐츠
research-pr	C0A7FS8UC66	public	리서치 PR
prompt	C0A98HXSVMK	public	프롬프트 엔지니어링/LLM 기법
deep-research-trending	C0AN34G4QHK	private	딥 리서치/논문 분석/트렌딩
idea	C0A6U3HE3GS	public	아이디어/브레인스토밍
random	C0A6CLTNARM	public	일반/기타
효정-할일	C0AA8NT4T8T	private	개인 태스크/할일
효정-주식	C0A7V1A09NK	private	주식/투자/트레이딩
효정-insight	C0A8SSPC9RU	private	전략 인사이트/분석
효정-이미지	C0A88DDE946	private	AI 이미지/비디오/음악 생성
ai-design	C0B1S9NT57B	public	AI 디자인/이미지 생성/크리에이티브

Step 3b: MCP Search (fallback)

If the channel name is not in the registry, search via MCP:

• Server: plugin-slack-slack
• Tool: slack_search_channels
• Query: the channel name the user provided (strip # prefix if present)

Extract the channel_id from the search results. If multiple matches, pick the closest exact name match.

Step 3c: Private channel fallback

If slack_search_channels returns no exact match, the channel may be private. Use slack_search_public_and_private to find it:

• Server: plugin-slack-slack
• Tool: slack_search_public_and_private
• Arguments: query: "in:{channel_name}", channel_types: "private_channel", limit: 1, response_format: "detailed"

The response will contain Channel: #channel-name (ID: C0XXXXXXXX). Extract the channel_id from this.

If the user did not specify a channel, or all lookup steps fail to find a channel, default to #press (C0A7NCP33LG). Do NOT ask the user to clarify — post to #press and note the fallback in the thread.

Step 3.5: Pre-Post Quality Gate

Before posting to Slack, verify all 3 thread messages are ready:

• Message 1 (Title): Bold Korean title with topic emoji, one-line summary, and direct tweet URL
• Message 2 (Summary): Contains Korean summary, 3+ key insights from web research, source links
• Message 3 (AI GPU Cloud): Contains ThakiCloud relevance analysis OR explicit "관련 없음" statement
• All messages use Slack mrkdwn (not markdown) — no ** or ##
• Media attachments prepared via 3-step Slack upload flow if original tweet has images/video

Media Upload Flow (MANDATORY when tweet has images/video)

If the tweet contains media.photos or media.videos, upload each media file to the Slack thread using scripts/slack_upload_file.py. This replaces unreliable URL unfurling with direct file uploads via Slack's files.uploadV2 API.

When to upload: After Message 1 is posted (to obtain message_ts for thread targeting), before or alongside Messages 2–3.

Single image/video:

python3 scripts/slack_upload_file.py \
    --url "{media_url}" \
    --channel {channel_id} \
    --thread-ts "{message_ts}"

Multiple media (batch mode):

python3 scripts/slack_upload_file.py \
    --urls "{url_1}" "{url_2}" "{url_3}" \
    --channel {channel_id} \
    --thread-ts "{message_ts}"

How it works internally (3-step files.uploadV2 flow):

1. files.getUploadURLExternal — request a temporary upload URL and file_id from Slack
2. HTTP POST to the temp URL — push the downloaded media bytes
3. files.completeUploadExternal — finalize the upload and share to channel_id + thread_ts

Media URL sources from FxTwitter API response:

• Photos: tweet.media.photos[].url (e.g. https://pbs.twimg.com/media/xxx.jpg)
• Videos: tweet.media.videos[].url (direct MP4 URL)

Error handling: If upload fails, log the error and continue with the remaining thread messages. Do NOT skip the entire post because of a media upload failure — the text content is still valuable.

If web research returned no relevant results, note this in Message 2 rather than omitting the section. See assets/templates/slack-thread.md for the 3-message template structure.

Step 4: Post to Slack (3 Messages)

All messages use Slack mrkdwn format. Rules:

• Use *bold* (single asterisk), _italic_ (underscore)
• Do NOT use **double asterisks** or ## headers
• Write content in Korean
• Limit each message to under 4000 characters

FORMATTING RULES — VIOLATING ANY OF THESE IS A QUALITY FAILURE:

0. MESSAGE 1 IS PLAIN TEXT ONLY — exactly 2 lines: one Korean headline sentence + source URL. NO emojis, NO bold, NO italic, NO author info, NO engagement stats, NO "원문 보기", NO flag emojis. Author/engagement data belongs ONLY in Message 2.
1. NO decorative emojis in section headers or body text (Messages 2 and 3).
   • ALLOWED emojis (exhaustive list, Message 2/3 only): ❤️ 🔁 👀 📎 1️⃣ 2️⃣ 3️⃣ (engagement stats and thread numbering only)
   • FORBIDDEN everywhere: 🔍 💡 🚀 📊 🎯 ✅ ⚡ 🔗 📌 💰 🏆 🧠 👤 📝 or ANY other decorative emoji
   • Flag emojis (🇮🇹 🇺🇸 🇰🇷 etc.) are FORBIDDEN everywhere
2. ALL body text MUST be in Korean. English is allowed ONLY for:
   • Proper nouns (product names, person names, company names)
   • Technical terms with no standard Korean translation
   • URLs and code snippets
3. Section headers (Messages 2 and 3 only) use bold text ONLY — no emojis before or after. Correct: 핵심 내용 Wrong: 🔍 핵심 내용 or 💡 Key Insights
4. Do NOT invent new section headers. Use ONLY the headers defined in the template: Tweet 요약, 핵심 내용, 인용 트윗, 추가 조사 결과, 참고 링크, AI GPU Cloud 서비스 인사이트, 핵심 시사점, 적용 가능성, Action Items, {주제} 인사이트
5. Do NOT add decorative separators (═══, ───, *** etc.)
6. Media upload is MANDATORY — if the tweet has photos or videos, upload them. HARD FAILURE if skipped.

Message 1: Title (Channel Post)

Post to the channel via Shell using scripts/slack_post_message.py (sends as user identity via SLACK_USER_TOKEN, NOT the RandomGame Slack app):

python3 scripts/slack_post_message.py \
    --channel "{channel_id}" \
    --message "{message_text}"

The script outputs JSON: {"ok": true, "ts": "...", "channel": "..."}. Parse ts as message_ts.

Format — Message 1 is EXACTLY 2-3 lines, nothing more:

{Korean headline — one sentence summarizing the core insight or news}
{source URL}

Message 1 MUST contain ONLY:

1. A concise Korean headline (one sentence, no emoji prefix, no bold markup)
2. The original tweet URL

Message 1 MUST NOT contain (these belong in Message 2):

• Author info (@username, name, bio, follower count) — NEVER in Message 1
• Engagement stats (likes, retweets, views, bookmarks) — NEVER in Message 1
• "원문 보기" link text — NEVER in Message 1
• Flag emojis (🇮🇹 🇺🇸 🇰🇷 etc.) — NEVER in Message 1
• Decorative emojis (🧠 👤 🔗 📝 🔍 💡 🚀 etc.) — NEVER in Message 1
• Multiple lines of description — NEVER in Message 1
• Bold/italic formatting — NEVER in Message 1

Anti-pattern (HARD FAILURE — must never produce this):

🧠 Claude Code + Obsidian = 세컨드 브레인 구축 완전 가이드
👤 @defileo | DeFi & AI 인사이트 크리에이터 (팔로워 41.7K)
🔗 원문 보기
🇮🇹 👀 365만 | ❤️ 5,331 | 🔄 429 | 📝 26,583

Correct example:

Claude Code + Obsidian으로 세컨드 브레인 구축하는 완전 가이드 — Karpathy LLM Wiki 기반 지식 관리 시스템
https://x.com/defileo/status/2042241063612502162

CRITICAL: Capture the message_ts (timestamp) from the response. This is needed for thread replies.

Message 2: Detailed Summary (Thread Reply)

Post as a thread reply via Shell using scripts/slack_post_message.py with --thread-ts:

python3 scripts/slack_post_message.py \
    --channel "{channel_id}" \
    --message "{message_text}" \
    --thread-ts "{message_ts}"

Format:

*Tweet 요약*
- 작성자: @{screen_name} ({name}) — {author bio/description}
- 반응: ❤️ {likes} | 🔁 {retweets} | 👀 {views}
- 작성일: {created_at}

*핵심 내용*
{Tweet 내용을 한국어로 요약 정리. 원문이 영어인 경우 번역 포함.}

{인용 트윗이 있는 경우:}
*인용 트윗 (@{quote.author.screen_name})*
{quote.text 요약}

*추가 조사 결과*
{WebSearch로 수집한 관련 정보를 bullet point로 정리}
- {관련 배경 정보}
- {최근 동향}
- {영향 및 의미}

*참고 링크*
- {검색에서 발견한 관련 URL들}

Message 3: AI GPU Cloud Insights (Thread Reply)

Post as another thread reply via scripts/slack_post_message.py with the same --thread-ts.

Format:

*AI GPU Cloud 서비스 인사이트*

{이 트윗 주제가 AI GPU Cloud / AI 플랫폼 서비스에 어떤 의미를 가지는지 분석}

*핵심 시사점*
- {GPU 클라우드 인프라 관점에서의 인사이트}
- {AI 플랫폼 서비스에 미칠 영향}
- {팀이 취해야 할 액션 또는 고려사항}

*적용 가능성*
{구체적으로 우리 서비스에 어떻게 적용하거나 대응할 수 있는지}

Message 3B: Topic-specific (Thread Reply) — when NOT AI GPU Cloud related

If the tweet topic is not directly related to AI GPU Cloud / AI platform services, use this topic-specific template instead of Message 3A:

*{주제} 인사이트*

{이 트윗 주제에 대한 심층 분석 — 시장 영향, 기술적 의미, 전략적 시사점 등}

*핵심 시사점*
- {해당 분야/주제 관점에서의 인사이트}
- {산업/시장에 미칠 영향}
- {관련 트렌드 및 향후 전망}

*Action Items*
- {구체적이고 실행 가능한 후속 조치}
- {모니터링해야 할 사항}
- {추가 리서치가 필요한 영역}

Step 5: Long-Form Detection → x-to-notion (optional)

After completing the Slack thread, evaluate whether the tweet qualifies as long-form content worth archiving to Notion.

Detection criteria (ANY match triggers Notion publishing):

Criterion	Threshold
Tweet text length	> 500 characters
X Article	tweet.article exists in FxTwitter response
Thread	Self-reply thread with 3+ tweets
Quote + body combined	Total text (tweet + quote) > 400 characters

Execution:

If any criterion matches, invoke the x-to-notion skill with the original tweet URL:

1. Log: "장문 콘텐츠 감지 — Notion 아카이브 시작 ({criterion})"
2. Run x-to-notion pipeline (Phase 1–5) with:
   • url: the original tweet URL
   • parent_page_id: default (3239eddc34e680e8a7a5d5b5eac18b38)
3. After x-to-notion completes, add a 4th Slack thread reply to the existing thread:

*Notion 아카이브*
장문 콘텐츠가 Notion에 전문 번역·보관되었습니다.
{Notion page URL}

Error handling: If x-to-notion fails, log the error but do NOT fail the overall x-to-slack pipeline. The Slack thread is already posted successfully.

Skip flag: If the user passes skip-notion, skip this step entirely.

Step 6: Artifact Save & URL Registration (MANDATORY)

After Slack posting completes (and optional x-to-notion), save the intelligence artifact to the research repo and register the URL.

Step 6a: Generate Markdown Artifact

Create a markdown file summarizing this tweet's intelligence:

---
url: {tweet_url}
author: @{screen_name}
date: {created_at}
slack_channel: {channel_name}
slack_ts: {message_ts}
topic: {classified_topic}
---
# {Korean title from Message 1}

{Message 2 content (full Korean summary with research)}

{Message 3 content (insights and action items)}

Save to a temporary file, e.g. /tmp/x-to-slack-{tweet_id}.md.

Step 6b: Save to Research Repo

python3 "scripts/intelligence/intel_registry.py" save \
  "{tweet_url}" "/tmp/x-to-slack-{tweet_id}.md" \
  --type tweet \
  --channel "{channel_name}" \
  --ts "{message_ts}" \
  --topic intelligence

This copies the artifact to knowledge-bases/intelligence/raw/ and registers the URL in the central dedup registry.

If intel_registry.py is not found, log a warning and skip (graceful degradation). The Slack post is still complete.

Examples

Example 1: English tech tweet

User says: "https://x.com/kaborogang/status/1893370063134626098 ai-gpu-cloud 채널에 올려줘"

Actions:

1. Fetch tweet via api.fxtwitter.com/kaborogang/status/1893370063134626098
2. WebSearch for key topics mentioned in the tweet
3. Find ai-gpu-cloud channel via slack_search_channels
4. Post 3-message thread (title, summary with research, GPU Cloud insights)

Result: Structured Korean-language thread posted to #ai-gpu-cloud

Example 2: Korean tweet with quote

User says: "x-to-slack https://x.com/user/status/123 #general"

Actions:

1. Fetch tweet — detect quote tweet in tweet.quote
2. WebSearch for topics from both the main tweet and quote
3. Find general channel
4. Message 2 includes quoted tweet summary section

Result: Thread with quote tweet context included

Error Handling

• FxTwitter API failure: Try Agent-Reach Twitter channel as fallback: twitter read {tweet_id} (requires TWITTER_COOKIE configured via agent-reach). If twitter-cli is also unavailable or fails, report error to user and do not post to Slack.
• Channel not found (public): Fall back to slack_search_public_and_private with in:{channel_name} and channel_types: "private_channel". If still not found, default to #press (C0A7NCP33LG).
• Missing thread_ts: If Message 1 response doesn't include message_ts, use slack_read_channel to find the most recent message just posted.
• Tweet has no text: Still process if media or quote exists; note empty text in summary.

MCP Tool Reference

Tool	Server	Purpose
slack_search_channels	plugin-slack-slack	Find public channel_id by name
slack_search_public_and_private	plugin-slack-slack	Fallback: find private channel_id via in:{name} query
scripts/slack_post_message.py	Shell (direct Slack API)	Post text messages and thread replies as user identity via SLACK_USER_TOKEN
slack_read_channel	plugin-slack-slack	Fallback to find message_ts
scripts/slack_upload_file.py	Shell (direct Slack API)	Upload media files to Slack via files.uploadV2 3-step flow