Extracts structured knowledge from book chapters into per-category JSON files: facts, examples, metaphors, quotes, glossary. Uses LLM prompts on text chunks with page estimates.
From reading-assistantnpx claudepluginhub xiaolai/reading-assistant-for-claude --plugin reading-assistantsonnetOptimizes queries, indexes, caching, and architectures for SQL (PostgreSQL, MySQL), NoSQL (MongoDB, DynamoDB), and cloud databases. Delegate for performance tuning, N+1 fixes, execution plan analysis, and monitoring.
Expert DB admin for cloud-native databases on AWS/Azure/GCP/OCI. Masters IaC (Terraform), HA/DR, performance tuning, multi-cloud strategies, compliance, cost optimization. Delegate proactively for architecture, operations, reliability engineering.
Expert backend security coder for input validation, authentication, API security, injection prevention, CSRF protection, and secure headers. Delegate for security implementations and code reviews.
Extract structured knowledge from book chapters into per-category JSON files.
Read metadata.json for chapter list and word counts
Read config for: models.extractor, categories to extract (default: all)
For each non-empty chapter:
a. Read chapters/<slug>.json and its chunks
b. For each chunk, prompt the LLM to extract items in ALL categories using the JSON schemas from the core skill
c. Compute page_estimate for each item using the derivation from the core skill
d. Accumulate items per category
Write output files:
extractions/facts.json — array of Fact itemsextractions/examples.json — array of Example itemsextractions/metaphors.json — array of Metaphor/Analogy itemsextractions/quotes.json — array of Quote items (all with verified: false)extractions/glossary.json — array of Glossary Term itemsIf configured model is external, use provider script via Bash (see core skill for routing)
For each chunk, use this prompt structure:
Given this text from chapter {N} of "{book_title}":
{chunk_text}
Extract ALL of the following (use the exact JSON schemas):
1. Facts: claims, data points, statistics, research findings
2. Examples: case studies, anecdotes, illustrations
3. Metaphors: figurative language, analogies, comparisons
4. Quotes: memorable phrases, punch lines, aphorisms
5. Glossary: domain-specific terms with definitions
Return a JSON object: { "facts": [...], "examples": [...], "metaphors": [...], "quotes": [...], "glossary": [...] }