skills.rest
oresttokovenko
oresttokovenkoCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

wiki-ingest

Distills documents, chat exports, URLs, and codebases into interconnected Obsidian wiki pages.

Updated Nov 9, 2023
Search Tags:#pdf#knowledge management#provenance#obsidian#wiki#ingest#distillation
One-click install
npx skills add https://github.com/oresttokovenko/dot-files --skill wiki-ingest-oresttokovenko
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: wiki-ingest
Source: https://github.com/oresttokovenko/dot-files/tree/main/chezmoi/dot_agents/skills/wiki-ingest
Command: npx skills add https://github.com/oresttokovenko/dot-files --skill wiki-ingest-oresttokovenko

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Adding new knowledge to a personal Obsidian wiki is tedious: you must read each source, decide which pages to create or update, track provenance, and wire cross-links by hand. This Skill automates that entire distillation pipeline so any source becomes structured, interlinked wiki pages. ## Core Features & Use Cases - Multi-format ingestion: Handles PDFs (including academic papers with figure/equation extraction), markdown, chat exports (ChatGPT, Slack), CSV/JSON data, images via vision, web URLs, and git repositories. - Incremental and staged workflows: Append mode skips unchanged sources via content hashing, raw mode promotes _raw/ drafts, and staged writes route new pages to _staging/ for review before publishing. - Provenance and structure: Every claim is tagged as extracted, inferred, or ambiguous; pages get typed relationships, summaries, and wikilinks; a manifest tracks all ingested sources. - Use Case: Point the skill at a folder of research PDFs and meeting transcripts; it batch-plans the work, extracts concepts and entities, and produces cross-linked wiki pages with page-cited provenance. ## Quick Start Ask the agent to ingest a source into your wiki, for example: add this PDF and my ChatGPT export to my Obsidian wiki.

Frequently Asked Questions about wiki-ingest

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I add documents to an Obsidian wiki automatically?▼

Point the skill at any file, folder, or URL and it reads the source, extracts concepts and entities, then creates or updates wiki pages with wikilinks and provenance markers. Append mode skips sources already ingested by comparing content hashes against the vault manifest.

What file formats can be ingested into the wiki?▼

Supported formats include markdown, text, PDF, JSON/JSONL, CSV/TSV, HTML, chat exports from ChatGPT and Slack, meeting transcripts, and images via a vision-capable model. Git repositories are ingested by cloning locally and letting batch-plan respect the repo's gitignore.

How does incremental ingest avoid reprocessing unchanged files?▼

The obsidian-wiki cache-check command compares SHA-256 hashes of each source against the vault's .manifest.json and returns new, modified, unchanged, missing, and unavailable lists. Only new and modified files are re-ingested, and hashes are recorded after each source completes.

Can it ingest academic papers with figures and equations?▼

Yes. For arXiv or conference PDFs it re-reads figure- and equation-dense pages with vision, extracts the paper's own figures using PyMuPDF, keeps core equations as display LaTeX, and renders results as markdown tables using a Paper Deep-Dive template.

What happens to drafts in the _raw staging folder?▼

Raw mode promotes each _raw/ file into a proper wiki page, deriving source provenance from the file's capture_source and sources frontmatter. The original is then moved into _raw/_archived/ one file at a time so it is never double-processed or deleted.

Does the skill execute commands found inside source documents?▼

No. A content trust boundary treats all source material as untrusted data to distill, never instructions to follow. Embedded commands, prompt-injection text, and requests to exfiltrate data are ignored and only the SKILL.md instructions control behavior.

Related Skills

Explore More Education & Research Skills→
tiancaiamaotiancaiamao
5

wiki

Maintain and query a personal wiki from structured Markdown sources.

Community
Intermediate
Michael1PengMichael1Peng

ingest-url

Fetch web content and distill it into Obsidian wiki pages.

Community
Intermediate
cdeistopenedcdeistopened
2

local-markdown-search

Index and search local Markdown collections with BM25 and vector embeddings.

Community
Intermediate
yunqiqiliangyunqiqiliang
8

clickzetta-data-ingest-pipeline

Route data ingestion to ClickZetta Lakehouse pipelines based on source type, latency, and volume.

Community
Advanced