Skip to content

po4yka/llm-wiki-skills / skills/wiki-ingest

wiki-ingest

Verified

Convert trusted raw sources into reviewable LLM-Wiki pages. Use when adding PDFs, articles, transcripts, notes, repository docs, web captures, or other source material into a raw/wiki/schema vault while preserving provenance, ambiguity, links, index updates, and log entries.

Wiki Ingest

Goal

Convert raw source material into reviewable wiki updates while preserving provenance, uncertainty and human-owned synthesis.

When to use

  • A trusted raw source (PDF, article, transcript, notes, repo doc, web capture) should become reviewable wiki pages.
  • Sorted inbox material was approved for ingestion after wiki-triage.

Route unsorted captures to wiki-triage first. A filled input-to-output example ships in assets/worked-example.md.

Inputs

  • One or more files under raw/ or inbox/.
  • Optional user goal: why the source matters.
  • Existing wiki/index.md and wiki/log.md.
  • Existing AGENTS.md, CLAUDE.md and schemas if present.

Write permissions

Allowed:

  • create/update wiki/sources/*;
  • create/update wiki/entities/*;
  • create/update wiki/concepts/*;
  • create/update wiki/comparisons/* when explicitly useful;
  • append to wiki/log.md;
  • update wiki/index.md;
  • write low-confidence material to _agent/drafts/ or mark it review_required: true.

Not allowed unless explicitly requested:

  • edit raw/ files;
  • delete source material;
  • overwrite ## My synthesis or human-owned sections;
  • mark generated pages as verified;
  • perform bulk rewrites unrelated to the source.

Procedure

1. Read the schema layer

Inspect AGENTS.md, CLAUDE.md, _meta/taxonomy.md, _meta/schemas/* and wiki/index.md when present.

2. Analyze before writing

Produce a short analysis plan before editing:

## Ingest analysis

- Source path:
- Source type:
- Likely page type:
- Candidate entities:
- Candidate concepts:
- Possible existing pages to update:
- Ambiguities:
- Review risks:

3. Extract claims

Classify claims using the shared support vocabulary:

  • extracted: directly present in the source;
  • inferred: reasoned from the source;
  • ambiguous: plausible but unresolved, requires human review;
  • synthesis: editorial conclusion — human-owned, do not generate during ingest;
  • unsupported: no inspected source currently supports the claim;
  • conflicting: inspected sources disagree about the claim.

Do not collapse competing claims into a single conclusion. Search the vault for existing pages on the same entities or concepts and re-verify each extracted claim against the raw source before writing it, so newly generated pages do not silently contradict already-verified material.

4. Write source page

Recommended source page structure:

---
type: source
status: draft
source_paths: []
ai_generated: true
ai_confidence: 0.0
review_required: true
---

# Source title

## Abstract

## Key extracted claims

## Entities

## Concepts

## Inferred implications

## Ambiguities and caveats

## Open questions

## Links created

5. Update entity and concept pages

When updating existing pages:

  • preserve human-owned sections;
  • append or patch rather than rewrite entire pages;
  • keep source backlinks near the claims they support;
  • mark inferred claims explicitly.

6. Update navigation

Update wiki/index.md only with pages important for navigation.

Append to wiki/log.md:

## [YYYY-MM-DD] ingest | <source title>

- Raw source: <path>
- Pages created:
- Pages updated:
- Review required:
- Open questions:

Final report

End with:

## Ingest summary

- Created:
- Updated:
- Needs review:
- Ambiguous claims:
- Suggested next sources:

Output

Ingestion produces new or updated Markdown pages under wiki/sources/, wiki/entities/, wiki/concepts/, and optionally wiki/comparisons/, plus an updated wiki/index.md entry and a matching log line in wiki/log.md. Each run ends with the ## Ingest summary report shown above, listing created pages, updated pages, items needing review, and open questions.

Safety gates

Stop and ask for review if:

  • the source conflicts with existing verified pages;
  • the source is too large and would require truncation;
  • source provenance is missing;
  • confidence is low for high-impact claims;
  • edits would touch protected human sections.
rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file).Stars: 198kUpdated Claude CodeCodexUnverifiedInstall
doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.Stars: 198kUpdated Claude CodeCodexVerifiedInstall