capture
/capture
This skill should be used when an agent needs to ingest a source (URL, YouTube, GitHub, PDF, clipboard, session log) into the vault. Triggers when wiki-scribe or wiki-researcher needs to "capture", "save source", "intake", "file URL", or "hash and store".
- Author
- ignromanov ↗
- Repository
- llm-obsidian-wiki ↗
- Licence
- MIT
- Last repository change
- 28.05.2026
Assessment and security
- Files in this package
- 1
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 1
No signals detected. Repository changed on 28.05.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
This skill should be used when an agent needs to ingest a source (URL, YouTube, GitHub, PDF, clipboard, session log) into the vault. Triggers when wiki-scribe or wiki-researcher needs to "capture", "save source", "intake", "file URL", or "hash and store".
Triggers from the description
Original author description
This skill should be used when an agent needs to ingest a source (URL, YouTube, GitHub, PDF, clipboard, session log) into the vault. Triggers when wiki-scribe or wiki-researcher needs to "capture", "save source", "intake", "file URL", or "hash and store".
How it works
- Purpose
- Trigger phrases (agent perspective)
- Inputs
- Outputs
- Workflow
- Web→MD stack (for capture-url.sh)
- Quality gate
- Scripts used
- Anti-patterns
SKILL.md preview
---
name: capture
description: This skill should be used when an agent needs to ingest a source (URL, YouTube, GitHub, PDF, clipboard, session log) into the vault. Triggers when wiki-scribe or wiki-researcher needs to "capture", "save source", "intake", "file URL", or "hash and store".
---
# Capture — Universal Source Intake
## Purpose
Ingest one source into `raw/<type>/<slug>.<ext>` (immutable, SHA-256 hashed) and create `wiki/sources/<slug>.md` (source-summary with `key_claims[]`).
Used by wiki-scribe (primary, passive intake) and wiki-researcher (via research workflow).
## Trigger phrases (agent perspective)
- "capture this URL"
- "save this PDF"
- "intake source"
- "file clipboard"
- "hash and store this"
- "wiki-scribe should ingest X"
## Inputs
One of:
- URL (any https://)
- YouTube link
- GitHub link (issue / PR / repo)
- PDF (path or URL)
- Raw text (clipboard, session log, transcript)
## Outputs
Two files per capture:
1. `raw/<type>/<slug>.<ext>` — immutable, SHA-256 in frontmatter, full content
2. `wiki/sources/<slug>.md` — source-summary with `key_claims: [{quote, anchor, confidence}]`, `quality`, `captured_by`
## Workflow
1. **Detect source type** from URL or path:
- `*.pdf` → capture-pdf.sh
- `youtube.com/*` or `youtu.be/*` → capture-youtube.sh
- `github.com/*/issues/*` or `*/pull/*` → capture-github.sh
- `github.com/*/pulls` → capture-prs.sh
- text without URL → capture-text.sh
- other URL → capture-url.sh
2. **Invoke the appropriate script** with `--vault $VAULT_PATH` and `--source $INPUT`. Scripts handle SHA-256 and frontmatter.
3. **Extract `key_claims`** for source-summary:
- If source <500 words: preserve full content as single claim
- If 500–5000 words: top-5 verbatim quotes with anchors (paragraph index)
- If >5000 words: top-10 quotes with anchors
4. **Set `quality` field** in source-summary:
- `high` if extraction succeeded with >500 chars
- `medium` if 200-500 chars (paywall preview, partial)
- `low` if <200 chars or extraction errors
5. **Set `captured_by`** to invoking agent name (`wiki-scribe` or `wiki-researcher`).
## Web→MD stack (for capture-url.sh)
The script tries three extractors in order:
1. `defuddle` (Node CLI, MIT, local) — primary
2. `trafilatura` (Python CLI, Apache-2.0, local) — fallback if defuddle fails OR content <500 chars
3. `r.jina.ai` (cloud) — opt-in via env `WIKI_ALLOW_CLOUD=1`
If all three fail, exit 3 and tell user to provide content via `capture-text.sh` (manual clipboard).
## Quality gate
- Every output has SHA-256 in frontmatter
- `wiki/sources/<slug>.md` is in tier 4
- `key_claims` is never empty for sources >500 words (if empty, log warning)
- `captured_by` is set
## Scripts used
- `${CLAUDE_PLUGIN_ROOT}/scripts/capture-url.sh`
- `${CLAUDE_PLUGIN_ROOT}/scripts/capture-youtube.sh`
- `${CLAUDE_PLUGIN_ROOT}/scripts/capture-github.sh`
- `${CLAUDE_PLUGIN_ROOT}/scripts/capture-prs.sh`
- `${CLAUDE_PLUGIN_ROOT}/scripts/capture-pdf.sh`
- `${CLAUDE_PLUGIN_ROOT}/scripts/capture-text.sh`
- `${CLAUDE_PLUGIN_ROOT}/scripts/capture-git-log.sh`
## Anti-patterns
- DO NOT interpret content during capture (preserve verbatim)
- DO NOT synthesize across sources (that's research's job)
- DO NOT edit existing wiki pages (only create raw/ + wiki/sources/)
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · ignromanov/llm-obsidian-wiki | 90 | MIT |
| wenlan-app/wenlan | 100 | Apache-2.0 |
| wenlan-app/wenlan | 100 | Apache-2.0 |
| optimiziramsi/skills | 100 | MIT |
| Kolibee-GmbH/m2b | 100 | MIT |
| davepoon/buildwithclaude | 100 | MIT |
| aw3-technology/vibehub-plugins | 100 | MIT |
| anurag-rivora/product-brain-starter | 100 | MIT |
| borghei/Claude-Skills | 100 | MIT |
Database snapshot: 13.11.2026 · package version e9c0b940a29f (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 90llmwiki-ingest
Ingest one source document (or a folder of them) into the llmwiki. Use when the user drops a new markdown file, PDF, or URL into the wiki and asks you to process it. The user will typically say "ingest this", "add this to the wiki", "process this file into the wiki", or point at a file under `raw/`. - 100ingest
Ingest one source (PDF, web page, or local markdown/text file) into the vault's LLM wiki — capture it into raw/, read it at the right cost, discuss it with the user, then write source/concept pages, update index, log and hot cache, and verify. Use when the user asks to ingest,… - 100wiki-ingest
ALWAYS use when the user wants a source saved to the vault — pastes an arXiv ID/URL, says 'add this to the vault', 'save this', 'remember this', 'I found this interesting', or describes a specific finding/technique/project to capture. Do NOT activate for general topic… - 90sessionwiki-ingest
Ingest one source document (or a folder of them) into the sessionwiki. Use when the user drops a new markdown file, PDF, or URL into the wiki and asks you to process it. The user will typically say "ingest this", "add this to the wiki", "process this file into the wiki", or… - 100wiki-ingest
Use when adding a new source to a wiki — a paper, article, URL, file, transcript, or any document. One ingest may touch 10-15 wiki pages. - 100mindbase
Maintain a persistent AI-tended research wiki. Activates when the user wants to ingest a source (PDF, URL, article, paste), capture a thought, build the wiki context, query what's already known, lint for orphans/contradictions, or maintain page conventions. Phrases include "add…