content-forge/.claude/skills/my-world/SKILL.md

name

description

my-world

One-shot context loader for the content-forge Obsidian vault. Loads projects, topics, drafts, knowledge graph, and current priorities into context. USE WHEN: User wants to see what's in the vault, check pipeline status, or start a session with full context awareness. Trigger phrases: 加载上下文、 vault 状态、我的世界、what's in my vault、load context、show pipeline。 DON'T USE WHEN: - User wants to CREATE new content → use write-article instead. - User wants to ANALYZE existing content deeply → use dao-analysis instead. - User wants to REVIEW a draft → use multi-review instead. - User wants to IMPORT an external file → use import-draft instead. - User says "写文章"、"分析"、"审查" → not this skill (read-only context loading only).

/my-world — Load My World

Load the entire content-forge vault context into the current session in one shot.

路由规则（Use when / Don't use when）

Use this skill when:

• Session 开始时需要加载 vault 全量上下文
• 需要查看当前 pipeline 状态（哪些选题/草稿/审核中）
• 需要了解知识图谱状态（orphans, dead ends, unresolved links）
• 用户说"加载上下文"、"vault 状态"、"我有什么在做"

Do NOT use this skill when:

• 用户想写新文章 → use write-article
• 用户想分析/拆解内容 → use dao-analysis
• 用户想审查草稿质量 → use multi-review
• 用户想导入外部文件 → use import-draft

Edge cases:

• "帮我看看 vault 里有什么" → this skill（查看 = context loading）
• "写一篇关于 XX 的文章" → not this skill（创作 = write-article）
• "分析一下我的草稿" → not this skill（分析 = dao-analysis 或 multi-review）

What It Does

1. Read vault stats (file count, tags, recent activity)
2. Scan all active content (topics, drafts, in-review)
3. Map the knowledge graph (backlinks, wikilinks, orphans)
4. Surface current priorities and blockers
5. Present a concise world-view summary

Important: Vault Directory

All obsidian CLI commands must run from within the vault directory to auto-detect the vault:

cd /home/kang/apps/content-forge/content-forge

Run all subsequent commands after this cd. The vault= parameter is unreliable; directory-based detection is the correct approach.

Execution Steps

Step 1: Vault Health

cd /home/kang/apps/content-forge/content-forge
obsidian vault
obsidian files total
obsidian tags counts sort=count

Step 2: Active Content Pipeline

Scan each pipeline stage for active items:

obsidian files folder="00-inbox"
obsidian files folder="01-topics"
obsidian files folder="02-drafts"
obsidian files folder="03-review"
obsidian files folder="04-published"

For each file found, read its frontmatter to extract title, status, content_type, channels:

obsidian properties path="<file-path>"

Step 3: Knowledge Graph

obsidian orphans
obsidian deadends
obsidian unresolved

Step 4: Recent Activity

obsidian recents

Step 5: Summarize

Present a concise summary in this format:

=== My World ===

Vault: content-forge (X files, Y tags)

Pipeline:
  inbox:     N items
  topics:    N items (list titles)
  drafts:    N items (list titles + status)
  review:    N items (list titles)
  published: N items total

Knowledge Graph:
  Orphans: N (notes with no incoming links)
  Dead ends: N (notes with no outgoing links)
  Unresolved: N (broken [[wikilinks]])

Recent: (last 5 files touched)

Priorities:
  - (inferred from draft/review status and recency)

Output

A structured world-view that gives full context for the session. No files are written — this is read-only.