PDF Chunk

Context-safe chunking for large/small-model docs (pdf-chunk)

Context-safe chunking for large/small-model docs

Part of the Universal PDF Management & Creation Skill Repository. Category: PDF Management. Every skill is individually callable (INDIVIDUAL MODE) and may also be invoked automatically by the orchestrator (SYSTEM MODE).

When to use

Prepare any document for retrieval without ever filling the context window.

Operations

• chunk by page
• chunk by max chars w/ overlap
• persist chunks + page text
• only the needed chunks are ever passed to the model

Engine (what actually executes this)

universal_pdf/core.chunk:chunk_document / chunk_text

All heavy lifting is done on disk / one page at a time by the shared engine — never by loading the document into an LLM context.

Tools & graceful degradation

• Preferred tool: local chunker (page & word boundaries)
• Fallback: —

If the preferred backend is unavailable the skill transparently uses the fallback or reports a documented limitation. It never reports success when an operation did not actually happen.

Validation (this skill attaches checks to its Outcome)

chunks>=1

Related skills

pdf-summarize, pdf-qa, pdf-index, pdf-provenance

Example requests it handles

• "Chunk this document for later Q&A."

Notes

Provenance: outputs cite their source page/section where relevant.