reading-assistant-for-claude
CLAUDE.md3 starsChanged 8 months agoMITView on GitHub
EPUB → structured knowledge pipeline for Claude Code. Commands orchestrate agents. Agents do the thinking. Scripts handle I/O and external providers. Pipeline: parse → chunk → [summarize + extract + review] → QC → [cards + index] scripts/llm_provider.py handles Claude/Ollama/OpenAI routing. Claude models → agent handles natively. External models → agent calls script via Bash. Never put AI logic in scripts — scripts only handle HTTP transport. All intermediate artifacts have formal JSON schemas in the design spec. Key files:…
- Reads credentials
In this repository2 other files it gives its agents.
- coreSkill
- obsidian-zettelkastenSkill
The file
We split this CLAUDE.md into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 21
Never put AI logic in scripts — scripts only handle HTTP transport.
- Very specific: line 25
All intermediate artifacts have formal JSON schemas in the design spec.
- Very specific: line 40
Run /read:init before first use
≈ “Common” is an estimate: at least 1 in 5 of the 27,641 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
reading-assistant
EPUB → structured knowledge pipeline for Claude Code.
Architecture
Commands orchestrate agents. Agents do the thinking. Scripts handle I/O and external providers.
Pipeline: parse → chunk → [summarize + extract + review] → QC → [cards + index]
- Phase 1 (batch): parser (+ language detection), chunker, review-fetcher
- Cost gate: estimate tokens, confirm with user
- Phase 2 (AI): summarizer, extractor, reviewer — configurable models via provider abstraction
- Phase 3 (QC): 4 parallel sub-agents, auto-retry, threshold gating
- Phase 4 (output): cardmaker (Obsidian), indexer (sqlite-vec RAG)
Provider Abstraction
scripts/llm_provider.py handles Claude/Ollama/OpenAI routing.
Claude models → agent handles natively. External models → agent calls script via Bash.
Never put AI logic in scripts — scripts only handle HTTP transport.
Data Contracts
All intermediate artifacts have formal JSON schemas in the design spec. Key files: metadata.json, chapters/.json, reviews/raw/.json, pipeline.json.
Embedding Provider Pinning
Each RAG database is pinned to one embedding provider (stored in _meta table). Switching providers requires --rebuild.
Config
User config: .claude/reading-assistant.local.md (YAML frontmatter)
Prerequisites
- Python 3.10+
- Run
/read:initbefore first use - Optional: OPENAI_API_KEY, HARDCOVER_API_KEY, Ollama
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

