Ai Trend Curation

Curate and post AI trend tweets from X (Twitter) with quote suggestions.

Security Scan

The skill's runtime instructions require local Node scripts and the xurl CLI but no code or install steps are included in the bundle, so the package is incomplete and its behavior is unclear.

Purpose & Capability

The skill's stated purpose (curate/post AI tweets to Slack) is plausible, but the metadata and SKILL.md both reference scripts/ai_trends.js and the xurl CLI as required runtime components — yet the registry entry contains no code files or install instructions. That mismatch means the skill cannot perform its stated tasks as provided, and it's unclear where the referenced scripts would come from.

Instruction Scope

SKILL.md explicitly instructs the agent to run node scripts (search, check-recent, format-blocks, mark-posted) and to call 'openclaw message send'. Those instructions are reasonable for the stated purpose, but they depend entirely on missing scripts. There are no instructions that request unrelated system files or extra environment variables, but the missing scripts create ambiguity about what other actions (network calls, credential usage, file I/O) those scripts would perform.

Install Mechanism

No install spec is provided (lowest install risk). However, because there is no install and no included scripts, the skill as-published is non-functional. If an operator later supplies or downloads the referenced scripts from an external source, that download/install step would become the primary risk and must be reviewed.

Credentials

The skill declares no required environment variables or credentials. Posting to Slack is done via the agent's 'openclaw message send' (which uses the environment the agent already has). No unrelated secrets are requested in the manifest or SKILL.md.

Persistence & Privilege

Default privileges (always: false, agent-invocable allowed) are used. The skill does not request permanent presence or attempt to change other skills' configs in the provided materials.

What to consider before installing

Do not install or enable this skill yet. The SKILL.md and metadata expect a Node script (scripts/ai_trends.js) and the xurl CLI, but no such script or install instructions are included in the package — this makes the skill non-functional and ambiguous about what it would actually run. Before proceeding, ask the author for: (1) the repository or source of scripts so you can review them, (2) an install spec or signed release (so you know binaries won't be fetched from untrusted URLs), and (3) details on how it authenticates to X (Twitter) and how Slack posting is authorized. If you obtain the scripts, review them for network endpoints, credential usage, and file I/O (look for code that reads environment variables, shells out to arbitrary commands, or posts to unknown external URLs). Also verify the xurl binary is the legitimate tool you expect. Only install after you can inspect the referenced scripts or are given a trusted upstream source; until then the skill should be considered incomplete and treated with caution.

X (Twitter) から AI 関連のトレンドツイートを収集・キュレーションし、Slack に投稿する。

ワークフロー

Step 1: ツイート検索

node scripts/ai_trends.js search

xurl CLI で日本語・英語の AI 関連ツイートを検索。いいね100以上（日本語）/ 500以上（英語）をフィルタ。

Step 2: キュレーション（エージェント実行）

検索結果をもとに、エージェントが以下を判断:

• 5-8件を選定（日本語3-5件 + 英語3-5件のバランス）
• 驚き屋・煽り系は排除。実務者・技術者の実践ベース投稿を優先
• 各ツイートに category, author_desc, text_ja(英語の場合), quote_suggestion を付与

Step 3: 重複チェック

node scripts/ai_trends.js check-recent

最近投稿済みのURLと重複しないか確認。

Step 4: Block Kit フォーマット

node scripts/ai_trends.js format-blocks \
  --tweets='[{...curated tweets...}]' \
  --summary="今日のAIトレンド要約"

Step 5: Slack 投稿

openclaw message send で親メッセージ + スレッド返信を投稿。

Step 6: 投稿済みマーク

node scripts/ai_trends.js mark-posted --urls='["url1","url2"]'

選定フィルター

• @yusaku_0426 が引用して付加価値を出せるか
• 元ツイートの発信者がRTしてくれそうか（批判・揚げ足取りNG）

引用ツイート案のペルソナ

一人称「僕」、丁寧語 or 体言止め。情報整理型キュレーター。 5-100文字で簡潔に。「最高！」「同意です」のような薄いリアクションは絶対NG。

引用パターン (A-H)

• A: 要点抜粋+補足解説
• B: 注目ポイント指摘（「地味に重要なのは〜」）
• C: 活用報告（「試してみた」＋体験）
• D: 共感（短い一言で同意）
• E: 一行要約（本質をワンフレーズに凝縮）
• F: 問いかけ・展開（「これ〜にも応用できそう」）
• G: 比較・接続（別ツール/事例と比較）
• H: 数字・データ強調

カテゴリ

• :rocket: new_release
• :hammer_and_wrench: tools
• :microscope: research
• :thought_balloon: opinions
• :bulb: tips