Skip to content

SKILL.md

Latest commit

 

History

History
162 lines (121 loc) · 4.34 KB

SKILL.md

File metadata and controls

162 lines (121 loc) · 4.34 KB
Raw
name content-graph-builder
description Build and maintain a personal or business content knowledge graph from chats, notes, project files, public content signals, and recurring ideas. Use when someone wants to turn scattered materials into a Markdown + wikilink graph, find content gaps, discover surprising connections, create weekly graph sync reports, or convert graph nodes into content, courses, products, or skills.

Content Graph Builder

Purpose

Help a user turn scattered thinking into a usable content graph.

The graph is not a passive archive. It should help answer:

  • What are my current content boundaries?
  • Which ideas, products, topics, and platforms are connected?
  • Which ideas are repeated but not yet structured?
  • Which nodes are isolated, overloaded, or missing a next action?
  • Which distant nodes create a useful surprising connection?
  • What can become a post, article, mini-course, product, Skill, or private archive?

Best Fit

Use this skill when the user wants to:

  • build a personal content brain
  • make a knowledge graph from chats and notes
  • map business ideas, accounts, products, and content assets
  • review weekly content signals
  • find content gaps and next actions
  • convert a graph node into public content or a product workflow
  • create a Markdown wiki that works with LLM Wiki, Obsidian, or other wikilink-aware tools
  • create a lightweight visual graph JSON for the included static viewer

Not For

Do not present this as:

  • automated data extraction
  • automatic traffic growth
  • viral-copy cloning
  • guaranteed content strategy

Use language like:

  • content graph
  • public content signals
  • review
  • judgment
  • workflow
  • content brain
  • knowledge map
  • surprising connections
  • content gaps

Default Project Shape

If the user has no existing project, create this shape:

content-graph-project/
├── README.md
├── purpose.md
├── schema.md
├── raw/
│   └── sources/
└── wiki/
    ├── index.md
    ├── overview.md
    └── log.md

If the user asks for a visual graph, also create:

graph-data.json

Use the JSON shape in references/visual-viewer.md so it can be imported into viewer/graph-viewer.html.

Workflow

  1. Inspect the user's current context and files. Do not infer private facts that are not present in provided or local materials.
  2. Define graph purpose: content planning, business strategy, creator workflow, learning map, project map, or weekly review.
  3. Normalize source material into short source summaries under raw/sources/.
  4. Create or update wiki pages with frontmatter and [[wikilink]] links.
  5. Classify nodes with stable types: person, product, concept, platform, asset, gap, connection, deliverable, source, synthesis.
  6. Update wiki/index.md, wiki/overview.md, and wiki/log.md.
  7. Produce at least one of:
    • content gap
    • surprising connection
    • deliverable recommendation
    • weekly sync report
  8. If a visual graph is requested, export a viewer-ready JSON file with sanitized node labels, short descriptions, edge types, and next actions.

Node Quality Bar

Every important node should include:

  • core judgment
  • connected nodes
  • what it can become
  • next action

Prefer updating existing pages over creating duplicate concepts.

Weekly Sync

A useful weekly sync should name:

  • new source signals
  • node changes
  • content boundary shifts
  • content gaps
  • surprising connections
  • deliverables to create next

Output Contract

When building a graph project, output:

## Project Created
- path:
- graph purpose:
- source files used:

## Nodes Added Or Updated
- ...

## Content Gaps
- ...

## Surprising Connections
- ...

## Next Deliverables
- ...

References

  • references/wiki-schema.md
  • references/sync-workflow.md
  • references/public-expression.md
  • references/visual-viewer.md
  • examples/sample-project/
  • examples/sample-viewer-data.json
  • templates/graph-project-template/
  • viewer/graph-viewer.html

Public Sharing Boundary

When preparing a graph for public sharing, publish the reusable workflow and sample data, not private archives.

Remove:

  • private file paths
  • thread IDs
  • unpublished client, business, or personal details
  • raw diary or relationship material
  • internal delegation notes
  • claims about guaranteed growth, traffic, or viral outcomes