What problem does it solve?
This Skill prevents costly PDF extraction failures and data errors caused by stale project configuration (such as outdated paths or mismatched package managers documented in CLAUDE.md) that no longer match the actual repository state, and addresses common edge cases like multi-numbered documents and dense unstructured fact sheets that produce incorrect or unlabeled extraction results.
Core Features & Use Cases
- Pre-extraction environment sanity gate: Validates project paths, resolves the correct package manager and interpreter from lockfiles (uv.lock or poetry.lock), and verifies extraction library imports with concrete version checks to eliminate silent toolchain failures.
- Multi-numbered document handling: Guides users to build page-offset maps for documents with multiple numbering schemes (like board books or regulatory filings) to avoid citation errors from mismatched page references.
- Metadata label validation: Ensures critical metadata labels (basis type, as-of date, currency) are verified against on-page statements rather than inherited from sibling data, preventing silent errors in downstream financial transforms.
- Dense fact sheet parsing: Provides layout-aware parsing guidance for single-page dense financial fact sheets and unlabeled bar-chart data streams to correctly map extracted numbers to their intended columns and series.
- Use Case: A financial analyst extracting data from a regulatory filing or portfolio fact sheet in a Python project can use this Skill to ensure extraction accuracy, avoid citation errors, and validate that all metadata labels are correctly sourced before running downstream analysis.
Quick Start
Use the pdf-extras skill alongside the pdf skill to run the environment sanity gate and follow all extraction best practices before pulling text or tables from any PDF in your Python project.
