pdf-intake-checkpoint

Create an intake checkpoint for PDF update tasks so downstream edits use a frozen task identity and a clear continuation gate.

SKILL.md

GitHub

PDF Intake Checkpoint

Purpose

Create a small intake checkpoint before editing so the task stays consistent across steps. This prevents scope drift (for example, mixing draft files or stale instructions) and makes the next action explicit.

When to use

Use this at the start of a PDF update task where:

  • an input PDF and instruction text are provided,
  • output path is predetermined,
  • later steps must remain reviewable and reproducible.

Inputs

  • /root/input/input.pdf
  • /root/input/input.txt
  • intended output path: /root/output/output.pdf

Actions

  1. Confirm the two required inputs exist.
  2. Read input.txt once and capture high-level obligations (fill/update fields, redact if requested, date/signature requirements).
  3. Write workflow/intake_checkpoint.json with frozen identity fields:
    • task_type
    • primary_input_pdf
    • instruction_source
    • primary_output_pdf
    • checkpoint_status
  4. Write workflow/continuation_gate.json with only the immediate next obligation:
    • next_stage: pdf-working-scope
    • gate_reason: standardize approved working set before editing
    • proceed_condition: working set record exists

Required outputs

  • workflow/intake_checkpoint.json
  • workflow/continuation_gate.json

Notes

  • Keep this stage lightweight: identity freeze + next-step gate only.
  • Do not edit the PDF in this stage.
  • Downstream stages should treat this checkpoint as the canonical starting record.