agentleFS
2233admin2233admin/obsidian-vault-bridge/skills/vault-ingest/SKILL.md

vault-ingest

Skill0 starsChanged 7 months agoMITView on GitHub

Ingest external content and REWRITE existing vault pages with new context. Not just "add a note" -- one source touches 5-15 existing pages. Resolves contradictions, updates claims, creates synthesis pages. The vault grows by integration, not accumulation.

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read the target note FIRST

    line 55
  2. Distinctive:

    Create synthesis page: if the source bridges 2+ existing domains.

    line 60
  3. Distinctive:

    After research returns results worth persisting

    line 91

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

/vault-ingest [URL or file path or pasted content]

Ingest a source and rewrite the vault to integrate it.

Steps

  1. Read _CLAUDE.md at vault root for folder map and conventions.
  2. Acquire the source:
    • URL -> fetch and convert to text
    • File path -> read the file
    • Pasted content -> use directly
    • Save raw source with frontmatter:
      ---
      date: YYYY-MM-DD
      tags: [research, ingested]
      source: [URL or path]
      content_hash: [first 8 chars of sha256]
      status: ingested
      ---
  3. Extract entities from the source:
    • Concepts, frameworks, mental models
    • Factual claims that could update or contradict existing notes
    • People, tools, libraries
    • Decisions, opinions, data points
  4. Map entities to vault -- spawn parallel search agents: For each entity, search the vault:
    • Does a note about this already exist?
    • Does the source agree with, update, or contradict what the vault says?
  5. Classify each entity's action:

    SituationAction
    New concept, no vault matchCreate note
    Existing note, source adds infoAppend new section
    Existing note, source contradictsMark contradiction with callout, cite both
    Existing note, source is newerUpdate claim, move old to "Historical" section
    Concept in 3+ notesConsider vault-graduate
  6. Execute rewrites -- spawn parallel write agents by folder, each must:
    • Read the target note FIRST
    • Find the right section (don't append at bottom blindly)
    • Preserve existing content, weave in new
    • Add source attribution
  7. Create synthesis page if the source bridges 2+ existing domains.
  8. Update: daily note, index.md, log.md
  9. Report:
    ## Ingestion Report: [source title]
    
    **Raw saved:** [[raw/source-name]]
    **Notes created:** [list]
    **Notes updated:** [list + what changed]
    **Contradictions found:** [list with both sides]
    **Synthesis pages:** [list]
    
    Total vault impact: [N] files touched

Rules

  • All output in the user's language (from _CLAUDE.md Language section). Translate ALL report headings: "## 摄入报告:", "已创建:", "已更新:", "发现矛盾:" for Chinese users. When ingesting content in a different language than the vault, translate key concepts while preserving original terms in parentheses. When updating existing notes, match that note's language.
  • Rewrite, don't just add -- integration, not accumulation
  • Never delete silently -- contradictions show both versions with dates
  • Raw sources are immutable -- edits happen in derived notes
  • content_hash for drift detection -- re-ingest detects if source changed
  • Attribution mandatory -- every claim cites its source
  • Search before create -- no duplicate notes

When to Suggest

  • User shares a URL or paper
  • User pastes a long block of content
  • After research returns results worth persisting

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.