All products

Organise PKA documents with AI

Review proposed destinations, retain extracted text beside originals and inspect a report of routed files, unavailable OCR and unresolved items.

Organise PKA documents with AI using the folder structure you already maintain. The pka-librarian skill inventories incoming files and proposes destinations for your local knowledge system. Without an existing repository map, it can infer likely routes from folder names and structure, with confirmation in standalone use.

For example, place several PDFs and a scanned image in an inbox and request proposals before anything moves. Available extraction tools can produce text sidecars without changing original content. Unavailable OCR is reported rather than silently counted as complete. Suspected transcripts are held for a separate decision.

What it does

  • Propose destinations for incoming documents
  • Extract text from PDFs when suitable tools exist
  • Report broken internal note links without automatic repair
  • Identify unresolved files in an inbox
  • Update an existing local search index
  • Surface stale references for human review

Install and use

  1. Load the complete skill folder and the applicable PKA conventions into an authorised file-and-command-capable agent. Text PDFs need pdfplumber or pypdf; scans need Tesseract and suitable page-to-image tooling.
  2. Choose the inbox, permitted destinations and confirmation mode. An inventory-only request can come first.
  3. Review routing proposals and confirm destinations, keeping uncertain files and transcripts separate.
  4. Inspect final paths, OCR limitations and the report. Request a non-destructive health check separately if wanted.

FAQ

Source

Made by rappdw / Dan Rapp. Original skill: pka-librarian, published on GitHub under the MIT licence.

View on GitHub