agent-harness-kindle-pipeline

Orchestrator for the Kindle Scribe pipeline. Subprocess-calls each agent-tool-* CLI in sequence per book listed in books.json. Source PDF -> Marker OCR -> figure scan -> strip pipeline (back-matter, marker-artifacts, clean-markdown) -> LLM proofread -> book-builder PDF + Obsidian export.

Required tools

This harness assumes the following tools are installed on PATH:

Tool	Repo
`clean-markdown`, `strip-back-matter`, `strip-marker-artifacts`	agent-tool-marker-cleanup
`scan-figures`	agent-tool-figure-scanner
`prose-quality`	agent-tool-prose-quality
`scrape-sources`	agent-tool-scrape-sources
`llm-proofreader`, `crop-flagged-blocks`, `reinsert-blocks`, `split-and-process`	agent-tool-llm-proofreader
`book-builder` (uses `agent-tool-pdf-builder`)	agent-tool-book-builder

Install all of them:

for repo in agent-tool-pdf-builder agent-tool-marker-cleanup agent-tool-figure-scanner agent-tool-prose-quality agent-tool-scrape-sources agent-tool-llm-proofreader agent-tool-book-builder; do
  git clone https://github.com/PatientVibes/$repo D:/$repo
  uv tool install --editable D:/$repo --reinstall
done
uv tool install --editable D:/agent-harness-kindle-pipeline --reinstall

Configuration

Copy books.example.json to books.json and fill in real source PDF paths:

cp books.example.json books.json  # then edit

Set OPENROUTER_API_KEY (used by the proofreader sub-tool):

export OPENROUTER_API_KEY='sk-or-...'

Usage

process-book <book-key-from-books.json>

Origin

Migrated from C:/Users/chris/OneDrive/Documents/Reading/tools/process_book.py (deleted post-migration). See the migration spec.