pdf

Files (2)

• SKILL.md
• task-routing.md · references/

SKILL.md (first lines)

# PDF

Use this skill when the task involves PDF files and success depends on extracting content reliably, generating PDFs, or validating rendered output.

## Read First

- the target PDF files
- `references/task-routing.md` to choose between lightweight text extraction and full rendering review
- `harness/state.md` if the task is part of an ongoing project workflow
- `harness/history.md` if it exists
## Goals

- choose the lightest PDF workflow that can still answer the user request
- avoid over-installing heavy dependencies for one-off text extraction tasks
- verify rendering visually when layout fidelity actually matters
- keep temp and output files organized

## Rules

- first classify the task as either `lightweight text extraction` or `full PDF review/generation`
- for resume screening, field extraction, keyword analysis, or quick content checks, prefer the lightweight path
- do not treat extracted text as proof of layout quality
- when the PDF is scanned, image-based, or visually important, route to the full rendering path
- prefer the user's or project's default Python environment before creating ad hoc environments
- only install system tools such as `poppler` when the task truly needs rendering or image export
- if dependency installation is blocked, say which dependency is missing and continue with the best fallback path that still preserves answer quality

## Process

1. Read `references/task-routing.md` and classify the task.
2. Run a quick environment preflight for Python and PDF tools.
3. If the task is text extraction, prefer `pypdf`, then `pdfplumber` if needed.
4. If the task is rendering-sensitive, use `pdftoppm`/`poppler` and inspect rendered pages.
5. For PDF generation, use stable programmatic tools such as `reportlab` and re-render for visual checks.
6. Keep intermediate artifacts in `tmp/pdfs/` and final repo outputs in `output/pdf/` when working inside a repository.
7. Report what workflow path was used, what dependencies were required, and what evidence supports the result.

## Output

Tell the user:

- which PDF workflow path was used: `lightweight text extraction` or `full PDF review/generation`
- which dependencies or tools were available, missing, or installed
- whether layout fidelity was actually verified visually
- where any output files were written
- any remaining evidence gaps such as scanned text quality or missing OCR

From gclm/super-stack, MIT.

More from gclm/super-stack