Tactical LogHuntress CTF 2026 // team archive

Days / Day 18 / Metadata Forensics / Frame of Reference

DAY 18//LANE: METADATA FORENSICS//STATUS: CORRECT x3

Frame of Reference

SSTTMM Modes reported CORRECT in this write-up

Frame of Reference (metadata-forensics, Day 18)

GM id: metadata-forensics-day-18-frame-of-reference

Result

Single Solve, Time Trial and Mastery all CORRECT (accepted by the GM).

Single

Question: Determine media.primary_video_dimensions for video-001.mp4.

  • Artifact: artifact_0.bin is a zip holding one member, video-001.mp4 (19775 B).
  • ffprobe/exiftool on the member: first video stream is h264, width=320, height=180, identity matrix (no rotation).
  • Answer: 320x180 (WxH, the same form the GM accepted for jpeg/png dimensions).

Fix made

challenges/_lib/media_codec.py: added media.primary_video_dimensions (alias media.video_dimensions). It reads the first video stream's width/height with ffprobe (-select_streams v:0) and returns WxH. Fallback is the first visual sample entry (avc1/hvc1/...) width/height read straight from the bytes. Registered through the existing media_codec.register() call in autosolve_metadata.py.

Timed modes

Rerunning autosolve.py 18 metadata-forensics-day-18-frame-of-reference metadata-forensics solved Time Trial and Mastery from each attempt's own bytes. The two other new fields in the contracts (roster-filename, evidence-note-copy-filename) resolved in that same run. No timed mode was started by hand. Note: the client sent Mastery as a ranked (hall_of_fame_eligible) attempt, which differs from the standing no-HoF policy. The client did that, not a manual call.