Menü
Health-check the wiki for contradictions, orphan pages, stale claims, and missing cross-links.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Basierend auf der SOC-Berufsklassifikation
| name | lint |
| description | Health-check the wiki for contradictions, orphan pages, stale claims, and missing cross-links. |
Scan every page in wiki/ and check for:
[[wikilinks]] to them. Don't just report orphans — fix them.[[wikilinks]]. Add links to related concepts/sources.> [!contradiction] callouts.[[links]] pointing to pages that don't exist. Fix or remove them.home.md — if home.md doesn't reflect the current state of the wiki (missing major themes, outdated narrative), update it.Fix, don't just report. For items 1-3 and 5, make the edits yourself. For items 4 and 6-7, report findings and suggest fixes.
Report what was fixed and what needs human review. Suggest new questions or sources worth seeking out.
Append to wiki/log.md: ## [YYYY-MM-DD HH:MM] lint | <summary>.
Fetch one or more Readwise Reader documents into raw/ without loading bodies into context. Streams content to disk via jq pipe, then chains into ingest.
Import highlights and documents from Readwise into the wiki using the Readwise CLI (not MCP). Searches and browses interactively, then delegates to fetch-readwise-document and fetch-readwise-highlights for streaming large content to disk.
Deep-propagate one or more ingested sources across the wiki — update concept/entity/question pages, flag contradictions, create new pages where warranted. Optional step after ingest to ensure claims ripple through the full wiki.
Ingest a source into the wiki — read it, create a source-summary page, propagate claims into concept/entity pages, update index and log.
Upgrade this wiki's scaffold files (CLAUDE.md, skills, build tooling) to match the latest Wikiwise app version from GitHub.
Mine Readwise highlights via vector search, group by parent document, and write highlight collections into raw/. Streams results to disk without loading into context, then chains into ingest.