Tactical LogHuntress CTF 2026 // team archive

Days / Day 06 / Warmups / Eight Is Enough

DAY 06//LANE: WARMUPS//STATUS: CORRECT x3

Eight Is Enough

Warmups SSTTMM Modes reported CORRECT in this write-up

Challenge

Category: warmups (lane warmups, GM id warmups-day-06). Day 06.

Artifacts

A single text blob delivered by the GM for this attempt (not saved to disk alongside the solver; the solver reads it directly from start's blob list).

Investigation

Reconstructed from the saved solver (scripts/solve.py) - no separate recon notes or digest were kept locally, so this writeup documents what the script itself does, which is the only surviving record of the method:

  • The blob is octal-encoded text: groups of octal digits separated by whitespace.
  • peel() decodes one octal layer at a time (int(x, 8) per whitespace-separated token), and the solver calls it in a loop (up to 10 times) until the data no longer matches the all-octal-digits pattern - i.e. it keeps peeling layers until it reaches plain text.
  • The final decoded text is split into lines; the last non-empty line is searched for a slug pattern word-word-word-NN (three lowercase hyphenated words followed by a two-digit number).
  • That matched slug is submitted verbatim as the answer for every answer_id the GM asked for.

Flag Recovery

The answer is the three-word-plus-number slug recovered by the regex above, after however many octal-decode passes were needed to reach it (the exact count is per-attempt, which is why the solver loops rather than hardcoding a fixed number of passes).

Minimal Reproduction

python3 challenges/_lib/ctf_gm.py  # provides GM()/solve_mode(), already on the path
python3 scripts/solve.py

Submits every mode (solve_mode handles the GM round-trip) and prints the result/remaining count.

Lessons Learned

  • "Eight" in the name plausibly hints at base-8 (octal) - worth checking a challenge's title for a numeric pun before assuming the encoding.
  • This challenge's writeup was not written at solve time, only the solver script was saved - the script itself turned out to be sufficient evidence to reconstruct an accurate writeup after the fact, which is the only reason this document exists. Writing the writeup alongside the solver, not after, remains the better practice.

Time Trial and Mastery (added 2026-10-11, writeup reconciliation)

All three modes are CORRECT on the Game Master (state.json: modes_done single_solve, time_trial, mastery). What the solve notes recorded for each mode (kb/lessons.jsonl, warmup_decode family):

  • Single Solve: one octal layer (3-digit groups); start, download, decode, submit.
  • Time Trial: the variant uses an atbash / rot style transform; the correct candidate is the one that produces the "Nice work" line, not simply the first candidate that decodes.
  • Mastery: 28 deliveries; each line is a fragment in one of decimal / octal / binary / hex / base64 / rot13. The lines are decoded per charset and joined; the deliveries are downloaded in parallel with one GM client per thread (a GM client socket must not be shared between threads).

Every timed attempt is a new variant, so the answers are derived from that attempt's own bytes each time. Not recorded at solve time: the wall-clock time of each timed mode. The solver script is the evidence.