agentleFS
dexio-wikidexio-wiki/llm-wiki-skills/skills/wiki-ingest/SKILL.md

wiki-ingest

Skill1 starsChanged 47 days agoMITView on GitHub

Compile a source (article, paper, transcript, meeting notes, doc, dataset) into an LLM wiki with provenance: keep an immutable raw copy with a hash, update every existing page the source touches instead of writing one summary page, cite the source at each claim, and label how far each claim can be trusted. Use when adding a new source to a wiki or when asked to read something and add it to the wiki.

45 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use wiki-status, wiki-research and 7 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • nexuschainsSkill

    wiki-status

    Reports the current state of the wiki vault — page counts, category breakdown, hub pages, orphans, stale entries, and a delta showing what has changed since the last ingest. Use this skill when the operator asks "what's the status", "wiki dashboard", "how many pages", "what's stale", "what's been ingested", "show me the delta", or "wiki health overview". Also surfaces structural insights (top hubs, fragmented clusters, bridge pages) when the operator asks "wiki insights", "what's central", or "show me the hubs". Read-only: this skill reports and recommends; it does not modify any files.

    Also in 5 projects that use this

  • Ivy-Chen1999Skill

    wiki-research

    Autonomously research a topic via multi-round web search, synthesize findings, and file structured results into the Obsidian wiki. Use this skill when the user says "/wiki-research [topic]", "research X", "find everything about Y", "do a deep dive on Z", "autonomous research on X", or wants comprehensive, web-sourced knowledge on a topic filed directly into their wiki.

    Also in 5 projects that use this

  • nexuschainsSkill

    cross-linker

    Scans the wiki vault and adds missing [[wikilinks]] between pages that are related but not yet connected. Use this skill when the operator says "link my pages", "find missing links", "cross-reference the wiki", "connect my wiki", "add wikilinks", or after any large ingest to weave new pages into the existing knowledge graph. Also trigger when orphaned pages need connecting or when wiki-lint reports fragmented tag clusters. This is a write-heavy skill — it modifies page files to insert links, unlike wiki-lint which only reports.

    Also in 4 projects that use this

  • bozhouDevSkill

    wiki-init

    Use when bootstrapping a new personal wiki for any knowledge domain — research, codebase documentation, reading notes, competitive analysis, or any long-term knowledge accumulation project.

    Also in 7 projects that use this

  • bozhouDevSkill

    wiki-update

    Use when revising existing wiki pages because knowledge has changed, a new piece of information updates or contradicts existing content, or the user wants to directly edit wiki content with LLM assistance.

    Also in 8 projects that use this

  • Ivy-Chen1999Skill

    claude-history-ingest

    Ingest Claude Code conversation history into the Obsidian wiki. Use this skill when the user wants to mine their past Claude conversations for knowledge, import their ~/.claude folder, extract insights from previous coding sessions, or says things like "process my Claude history", "add my conversations to the wiki", "what have I discussed with Claude before". Also triggers when the user mentions their .claude folder, Claude projects, session data, past conversation logs, local-agent-mode sessions, or audit logs.

    Also in 3 projects that use this

  • +1 more from nexuschains
  • +2 more from Ivy-Chen1999

The file

Files
SKILL.md

Ingest a source

Ingesting is where a wiki compounds or rots. Done well, one source updates the ten pages it bears on and every claim can be traced back to it. Done badly, it adds one more summary page that repeats the source, links to nothing, and hardens an unchecked claim into a "fact" that later pages cite.

Steps

  1. Keep the raw copy. Save the source as markdown under raw/ (for example raw/articles/2026-09-28-vendor-pricing-update.md) with this frontmatter:

    ---
    source_url: https://example.com/post
    ingested: YYYY-MM-DD
    sha256: <hex digest of the body after the closing --->
    ---

    Compute the hash over the body only (python3 -c "import hashlib,sys; print(hashlib.sha256(open(sys.argv[1],'rb').read().split(b'\n---\n',1)[1]).hexdigest())" file.md, or hash the text before you add frontmatter). Before saving, search raw/ for the same hash or URL. Same hash: already ingested, stop. Same URL, new hash: the source changed; save the new version beside the old one and note what changed. Never edit a raw file after ingestion.

  2. Orient (wiki-orient) on the source's main subjects.
  3. List the claims the source makes that matter to this wiki. For each, note whether it is new, confirms a page, updates a page, or contradicts a page.
  4. Update the pages the claims belong on (wiki-record, steps 3 to 7). One source often touches many pages: the entity it is about, the concepts it uses, a comparison it changes, a decision it bears on. Cite the raw file or URL at each claim.
  5. Label trust as you write, in the sentence that makes the claim:
    • vendor-sourced: a company's claim about itself (pricing, user counts, benchmarks).
    • self-reported: an individual's claim about their own results.
    • unverified: anything you recalled rather than read in this source.
    • estimate: a figure the source or you derived rather than measured.
    • The sampling frame, for any rate or percentage: who was counted, and who could not have been. "4 of 21 founders on the list" means nothing without what the list selected on.
  6. Handle contradictions with wiki-conflicts. Never let a newer source silently overwrite an older claim unless it is clearly the more authoritative record.
  7. Write a summary page only if the source is itself a subject: a landmark paper, a decision memo, a meeting that set direction. Otherwise the updated pages are the summary.
  8. Leave one change note per page naming the source, and tell the person which pages changed.

Batches

For many sources, ingest one at a time and finish each before the next. Parallel ingestion by several agents collides on the same shared pages; if you must parallelize, split the batch by subject so no two agents touch the same page, and follow wiki-shared.

Pitfalls

  • A summary page per source with nothing else changed. That is a document folder with extra steps; the value of the pattern is that existing pages get better.
  • Dropping the provenance during synthesis. A claim with no source attached will be repeated by the next agent as settled fact.
  • Quoting a vendor's number without the label. Vendor figures are hypotheses until checked against something the vendor does not control.
  • Ingesting what should stay out: secrets, private data, or content under terms that forbid copying. Summarize and link instead of storing the full text when in doubt.

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.