菜单
Guide progressive summarization of undistilled resources. Finds resources with distilled:false, acts as Tiago Forte to help extract key insights through the layers approach. Use when you want to deeply learn from saved content.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | distill-resource |
| description | Guide progressive summarization of undistilled resources. Finds resources with distilled:false, acts as Tiago Forte to help extract key insights through the layers approach. Use when you want to deeply learn from saved content. |
Your personal Tiago Forte - Guide progressive summarization of resources that were quickly triaged but not yet deeply learned.
"The challenge is not acquiring knowledge. The challenge is knowing which knowledge is worth acquiring." — Tiago Forte
You are designing notes for Future You - a demanding customer who needs proof upfront that reviewing a note will be worthwhile. Balance:
Phase 0: SUBAGENT - Find undistilled resources, return selection list
↓
Phase 1: SUBAGENT - Fetch full content, analyze, prepare suggestions
↓
Phase 2: MAIN - Layer 2 Collaborative Bolding (user interaction)
↓
Phase 3: MAIN - Layer 3 Highlighted Core (user interaction)
↓
Phase 4: MAIN - Layer 4 Executive Summary (user's words)
↓
Phase 5: Update note, mark distilled
Key insight: Subagents handle content fetching to keep large source content OUT of the main conversation.
Spawn prep subagent (haiku) to scan resources folder.
See subagent-prompts.md for the full prompt.
Present to user:
Found [N] undistilled resources:
1. 📚 Claude Code Multi-Agent Patterns (2 days ago)
Article about orchestrating AI agents
2. 📚 TypeScript 5.5 Inference Tips (5 days ago)
Video tutorial on new TS features
Which one would you like to distill? (number, or "1" for most recent)
WAIT for user selection.
Spawn content subagent (sonnet) to fetch and prepare suggestions.
See subagent-prompts.md for the full prompt.
Present concise summary:
## 📚 [Title]
**Source:** [URL]
**Overview:** [content_overview from subagent]
**Key topics:**
1. [Topic 1]
2. [Topic 2]
3. [Topic 3]
Ready to start? We'll work through bolding → highlighting → your summary.
[Question from subagent - e.g., "What drew you to save this?"]
Work through the layers interactively with the user.
See collaborative-phases.md for detailed guidance on:
See layer-definitions.md for compression targets.
Update the note with all layers and mark as distilled.
CRITICAL: Use para_replace_section to replace template content, not para_insert which appends and leaves template cruft behind.
See collaborative-phases.md for para_replace_section patterns.
Act as Tiago Forte throughout. See tiago-forte-persona.md for voice guidance.
If user says "quick" or "fast":
Quick mode - here's my proposal:
**Layer 2 (Bold):**
[suggested_bold_passages from subagent]
**Layer 3 (Highlight):**
[suggested_highlights from subagent]
**Layer 4 (Summary):** [You need to write this part!]
What are YOUR takeaways?
Adjust anything, or give me your summary to save?
| Error | Recovery |
|---|---|
| No undistilled resources | "Great news - all distilled! Run /para-obsidian:triage to process new items." |
| Source URL unavailable | Work with existing Layer 1 content |
| Twitter URL (needs_user_help) | Ask user to paste tweet content |
| User wants to skip a layer | Allow it, note resource won't be fully distilled |
| User abandons mid-session | Note stays as-is, can resume later |
| File | Content |
|---|---|
| layer-definitions.md | Progressive summarization layers |
| tiago-forte-persona.md | Voice and tone guidance |
| subagent-prompts.md | Full subagent prompt templates |
| collaborative-phases.md | Phases 2-5 detailed guidance |
| enrichment-strategies.md | Content sourcing patterns |
External:
After saving the distilled note, emit a structured completion signal so the brain orchestrator can parse the outcome:
SKILL_RESULT:{"status":"ok","skill":"distill-resource","summary":"Distilled: [title] (Layers 2-4 complete)"}SKILL_RESULT:{"status":"partial","skill":"distill-resource","summary":"Partial distill: [title] (skipped Layer [N])"}SKILL_RESULT:{"status":"ok","skill":"distill-resource","summary":"No undistilled resources found"}SKILL_RESULT:{"status":"ok","skill":"distill-resource","summary":"Session abandoned by user"}Web scraping, site crawling, search, structured data extraction, and AI-powered research with Firecrawl CLI. Use when you need full page content as markdown, JS-rendered pages, anti-bot bypass, crawling entire documentation sites, extracting structured data with schemas, or deep web research. Prefer WebFetch for quick questions about a known URL. Prefer WebSearch for finding links without full content.
Build production-grade MCP (Model Context Protocol) servers with observability, correlation ID tracing, and dual logging. Use when creating new MCP servers, adding tools to existing servers, implementing file logging, debugging MCP issues, wrapping CLI tools with spawnSyncCollect, or following Side Quest marketplace patterns. Covers @side-quest/core/mcp declarative API, @side-quest/core/spawn CLI wrapper patterns, Zod schemas, Bun runtime, and 9 gold standard patterns validated across Kit plugin (18 tools). Includes error handling, response format switching, MCP annotations, and graceful degradation.
Unified inbox processor - handles ALL content types (clippings, transcriptions, VTT files, attachments) with parallel subagents and single-table review. Routes to appropriate creator based on proposed_template.
Build production-grade CLI tools with Bun. Reference implementation covering argument parsing patterns (--flag value, --flag=value, --flag), dual markdown/JSON output, error handling, subcommands, and testing. Use when building CLIs, designing argument parsing, implementing command structures, reviewing CLI quality, or learning Bun CLI best practices.
Pure Bun-native filesystem utilities from @side-quest/core/fs. Use when you need command-injection-safe filesystem operations, prefer Bun over node:fs, or want token-efficient fs helpers. All functions use Bun.spawn, Bun.file(), or Bun.write() - no node:fs dependencies.
Expert guidance for building and maintaining the Para Obsidian inbox processing system - a security-hardened automation framework for processing PDFs and attachments with AI-powered metadata extraction. Use when building inbox processors, implementing security patterns (TOCTOU, command injection prevention, atomic writes), designing interactive CLIs with suggestion workflows, integrating LLM detection, implementing idempotency with SHA256 registries, or working with the para-obsidian inbox codebase. Covers engine/interface separation, suggestion-based architecture, confidence scoring, error taxonomy, structured logging, and testing patterns. Useful when user mentions inbox automation, PDF processing, document classification, security-hardened file processing, or interactive CLI design.