Skip to content

Add an audited local conversion MVP - #60

Draft
playchaodd68 wants to merge 4 commits into
firecrawl:mainfrom
playchaodd68:henry/mvp-20260807
Draft

Add an audited local conversion MVP#60
playchaodd68 wants to merge 4 commits into
firecrawl:mainfrom
playchaodd68:henry/mvp-20260807

Conversation

@playchaodd68

@playchaodd68 playchaodd68 commented Aug 7, 2026

Copy link
Copy Markdown

Summary

  • add an isolated henry-mvp wrapper for audited local document-to-Markdown conversion
  • report safe provenance and structural metrics without serializing paths or document content
  • surface mixed-PDF OCR and encoding warnings instead of reporting false-clean conversions
  • enforce private output modes and rollback-safe Markdown/report publication
  • pin a repository-local Rust toolchain and add a full-history privacy audit

Why

This turns the library into a reproducible local workflow for converting a confidential document while keeping the source and generated output outside Git. The wrapper also makes partial-conversion risks and output-boundary behavior explicit.

Impact and scope

  • the upstream library API remains available as before
  • no private source PDF, extracted Markdown, or generated report is committed
  • image-only PDFs still fail with a typed OCR-required error
  • mixed PDFs retain extractable text and emit safe warnings

Validation

  • 208 upstream library tests, 1 robustness test, and 8 snapshot tests passed; 1 local-samples test ignored
  • 18 MVP integration tests passed
  • locked tests, rustfmt, and Clippy with warnings denied passed
  • the real 23-page local conversion completed with source/fixture hash equality
  • private directory/file modes and four-commit full-tree privacy audit passed

View with [code]smith Autofix with [code]smith
Need help on this PR? Tag @codesmith-bot with what you need. Autofix is disabled.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant