v1.0.3

Features

• Decouple vectorization from file ingest — phased ingest pipeline
• Per-entity status breakdown in fp status output
• Log individual file paths for inserts/updates in incremental mode
• MCP discoverability resources — status, projects, access-policies
• Skip oversized files during vectorization with end-of-run summary
• Pre-scan large files and directories before vectorization with approval

Fixes

• Filter removed messages from MCP status count
• Cap MCP search limit at 200 to prevent protocol overflow
• Recursive folder file count via path prefix instead of unpopulated FK
• Default max_file_size_mb to 50MB in config template

Other Changes

• Stop hardcoding schema DEFAULTs in INSERT paths
• Remove dead _vectorize_file code from FileIndexer
• Replace non-existent fp assign with fp upsert in data model docs
• Resolve all 63 ruff lint errors across codebase
• Update all URLs to harringjohn/footprinter-cli
• Fix stale script references and CLI polish