Context
The v0.6.1 meeting document parser already enforces a 50 MiB input limit and counts actual decompressed DOCX bytes with a 64 MiB early-stop limit. Independent review found no release-blocking issue.
Follow-up
Add a defensive upper bound for the number of ZIP entries processed before iterating DOCX contents. This limits central-directory amplification from archives containing a very large number of empty entries.
Scope
- Add a deterministic ZIP entry-count limit and user-facing error.
- Add a synthetic regression fixture; do not use real meeting files.
- Keep the existing actual decompressed-byte limit.
This is a hardening follow-up, not a v0.6.1 blocker.
Context
The v0.6.1 meeting document parser already enforces a 50 MiB input limit and counts actual decompressed DOCX bytes with a 64 MiB early-stop limit. Independent review found no release-blocking issue.
Follow-up
Add a defensive upper bound for the number of ZIP entries processed before iterating DOCX contents. This limits central-directory amplification from archives containing a very large number of empty entries.
Scope
This is a hardening follow-up, not a v0.6.1 blocker.