Project Is
The Curator is a local Node.js web application that ingests text sources (PDF, MD, TXT) and automatically builds an interconnected knowledge wiki. The wiki is stored as plain
1022d agoReviewMountCreate a free account to mount
implementation is done.
Tier 2 in this repo:
- **OCR routing** — what sends a PDF down the Surya path, the silent-`fitz`-failure detection,
`--force-surya`. It only exists
1003mo agoReviewMountCreate a free account to mount
Schema and harness document for the CMDS LLM Wiki vault. Defines the 3-layer architecture (Raw Sources / Wiki / Schema), ingest-query-lint operations, file conventions, and frontmatter standards. This is the single source of truth for LLM behavior in this vault.
10012d agoReviewMountCreate a free account to mount
into a shareable one-page summary | ~1 min |
| `/autodecision:publish` | Ship a brief as PDF → Notion, email draft, gist, Slack, Drive, or Local | ~1 min |
| `/autodecision:plan` | Interactive setup wizard
986mo agoReviewMountCreate a free account to mount
GitHub repositories, remote URLs, and any file format supported by docling — DOCX, PPTX, XLSX, PDF, HTML, EPUB, images, and more) — extracts entities — delegates to `/bedrock:preserve` |
| `/bedrock:preserve` | Single write
966mo agoReviewMountCreate a free account to mount
/dashboard/docs`. Feature documentation lives here.
- `documents/` — Not documentation: screenshots for the README, the architecture PDF and its source, announcements.
- `shared/sql/README.md` — Database schema reference
Behavioral guidelines to reduce common LLM coding
9512d agoReviewMountCreate a free account to mount
prefer Mermaid (renders everywhere). Fallback to ASCII art
- Max line width: 100 characters (for PDF rendering)
### Length Guidelines
- Module total: **800-1500 words** (sweet spot for 20-40 min reading
957mo agoReviewMountCreate a free account to mount
citation graphs
- Hugging Face datasets search
- Web search and page crawling
- LaTeX compilation and PDF reading
- LLM-powered document QA and research proposal workflows
**Tech Stack:**
- Python 3.12+ with type
9413mo agoReviewMountCreate a free account to mount
module scripts. That is exactly how pdf.js ships its worker (`pdf.worker.min-*.mjs`), so every PDF preview ran without one in production while vite's dev server typed it correctly — invisible
8920d agoReviewMountCreate a free account to mount
directories use **noun/doer form** in kebab-case: `agent-skill-reviewer` not `agent-skill-review`, `pdf-translator` not `translate-pdf`.
- This aligns with skills being "personified" agent roles.
### Key design principle
875mo agoReviewMountCreate a free account to mount
python -m examples.spam_detection` - Run example
- `bachata-example` - Run spam detection example
- `bachata-pdf-example` - Run PDF extraction example
## Documentation
- Generate docs on each version
```bash
uv run pdoc
8616mo agoReviewMountCreate a free account to mount
A file Claude Code reads at the start of every session. It holds the commands, conventions and warnings the agent needs for this project.
Where does it go?
At the repository root. Claude Code also reads CLAUDE.md files in subdirectories when it works there.
What should it contain?
Build and test commands, the project's layout, conventions that aren't obvious from the code, and mistakes to avoid. Short files tend to work better than long ones.
CLAUDE.md or AGENTS.md?
Claude Code reads CLAUDE.md; most other agents read AGENTS.md. Many projects keep one and point the other at it.