laststance/newsletter-digest

Newsletter Digest — Deep Technical Summary

Codex Compatibility

When running this skill in Codex, translate Claude Code-only primitives before acting: AskUserQuestion -> chat/request_user_input, TodoWrite -> update_plan, Task/TaskCreate/TeamCreate/SendMessage -> spawn_agent/send_input/wait_agent when available and allowed, and EnterPlanMode/ExitPlanMode -> a concise chat plan plus explicit approval. Resolve Read/Write/Edit/Bash/WebSearch/WebFetch to Codex file/shell/web tools, and map ~/.claude/... paths to ~/.agents/... or ~/.codex/... unless the task explicitly targets Claude Code.

Cursor Compatibility

When running this skill in Cursor Agent, translate Claude Code-only primitives before acting: AskUserQuestion -> AskQuestion; TodoWrite -> Cursor TodoWrite or an equivalent checklist; Task/TaskCreate/TeamCreate/SendMessage/multi-agent flows -> Cursor Task (subagents), parallel Tasks, or run_in_background when allowed (TeamCreate/SendMessage may have no exact match); EnterPlanMode/ExitPlanMode -> Plan mode (SwitchMode / CreatePlan) plus explicit user approval. Resolve Read/Write/Edit/StrReplace/Bash/web/search/MCP via Cursor Composer or Agent equivalents. MCP names written as mcp__server__tool typically map to call_mcp_tool with configured server identifiers. Map ~/.claude/... to ~/.cursor/skills/, .cursor/skills/, and .cursor/rules/ unless the task explicitly targets Claude Code.

Reads a tech newsletter from Gmail and produces a comprehensive, structured summary with 5x the normal detail depth. Each article gets technical background, ecosystem impact, and contextual analysis.

<essential_principles>

Always Active

• 5x Detail Depth: Every article gets technical background, "why it matters", and ecosystem context — not just a one-line summary
• Adaptive Categorization: Match the newsletter's own section structure (headers, topics), don't force a fixed template
• Technical Accuracy: Use Context7 and Exa to verify claims and add supplementary technical details for main articles
• Structured Thinking: Use sequential-thinking to organize the analysis before writing output
• Language: Follow user's language preference. Default to Japanese per user config (respond in the language the user uses)
• Preserve Links: Keep all article URLs as markdown links for easy access

</essential_principles>

Phase 1: Retrieve Newsletter

1. Search Gmail using mcp__claude_ai_Gmail__gmail_search_messages:
   • If user provides exact subject → subject:"exact subject here"
   • If user provides newsletter name only → subject:"{name}" OR from:"{name}" with maxResults: 5, let user pick
   • If user says "latest" → search by newsletter name, sort by recent
2. Read the full message with mcp__claude_ai_Gmail__gmail_read_message
3. If the email body is truncated or too short, try gmail_read_thread for the full content

Phase 2: Analyze & Structure

Use mcp__sequential-thinking__sequentialthinking to:

1. Identify all articles/items in the newsletter — count them, note categories
2. Classify by section — detect the newsletter's own section headers (e.g., "IN BRIEF", "RELEASES", "ARTICLES", "CODE & TOOLS")
3. Prioritize main articles (typically 2-4 featured items with longer descriptions) vs brief mentions
4. Plan the output structure — map newsletter sections to output sections

Phase 3: Enrich Main Articles

For the top 2-4 featured/main articles:

1. Context7 (mcp__context7__resolve-library-id → mcp__context7__query-docs): Look up technical details for libraries/frameworks mentioned
2. Exa (mcp__exa__web_search_exa or mcp__exa__get_code_context_exa): Search for additional context, code examples, or related announcements
3. Add technical background that the newsletter's brief description omits

Phase 4: Generate Output

Follow the format in references/baseline-format.md.

Output Structure

---

## 📬 {Newsletter Name} #{Issue Number} — 詳細サマリー
**発行日: {Date} | 編集: {Editor} ({Publisher})**

---

## 🔶 メイン記事
[Each main article: 500-800 chars with technical depth]
[Include ★ Insight blocks after major articles]

## 📋 短信 (IN BRIEF)
[Each brief: 100-200 chars]

## 📦 リリース情報
[Each release: 50-150 chars with version + key changes]

## 📖 記事・動画
[Each article: 200-400 chars]

## 🛠 コード＆ツール
[Each tool: 100-300 chars]

## 🌐 エコシステム情報
[Each item: 100-300 chars]

---

★ Insight (号全体のテーマ分析)

Writing Rules

| Element | Rule | |---------|------| | Main articles | 500-800 chars each, include "why it matters", technical background, ecosystem impact | | Brief items | 100-200 chars, key takeaway only | | Releases | Version number + top 2-3 changes | | Tools/Libraries | What it does + why it's notable | | ★ Insight blocks | After main articles AND at the end (overall theme analysis) | | Sponsor content | Include with (SPONSOR) tag, keep brief | | Section headers | Use emoji + Japanese section names, adapt to newsletter's own categories | | Links | Preserve original URLs as markdown links |

Section Emoji Map

| Category | Emoji | |----------|-------| | Main/Featured | 🔶 | | Brief/Short | 📋 | | Releases | 📦 | | Articles/Tutorials | 📖 | | Code/Tools | 🛠 | | Ecosystem/Community | 🌐 | | Videos | 🎬 | | Opinions/Commentary | 💬 |

Adapt categories based on what the newsletter actually contains. Not every section needs to appear.

<success_criteria> A successful newsletter digest:

• [ ] All articles from the newsletter are covered (none skipped)
• [ ] Main articles have 5x depth with technical context
• [ ] Sequential-thinking was used to structure the analysis
• [ ] Context7/Exa enriched at least the main articles
• [ ] Output follows the baseline format structure
• [ ] ★ Insight block provides cross-cutting theme analysis
• [ ] All original URLs are preserved as markdown links </success_criteria>