Crystal Clear

The challenge

"Crystal compiles to native code with a distinctive shine. Fingerprint the runtime, then follow the DNS labels — this one has somewhere it wants to resolve."

130 questions in the Single Solve: 10 per day for Days 1 to 13. Earlier days' questions are the same as Day 12's. The 10 new d13_* fields have the same shapes as Day 12's:

Answer kinds across all 130: 66 exact text, 46 pattern, 9 integer, 5 JSON, 4 enum.

Reading the prompt: what it said we'd need

Safety

Static only. The evidence zips (≈32 MB each) were read in memory on our Kali VM. No binary was executed. This write-up was built from the solver's records, the saved Game Master replies, our record of ranked times, and the request log only. The zips weren't reopened for it.

What was in the files

One evidence.zip (32.2 MB) with one program per day. The solver's record shows these answers for the Day 13 file in the Single Solve set:

• format elf, architecture x86-64, machine 62: a 64-bit x86-64 ELF;
• config campaign BR13-000, domain node-000.day13.example.invalid, mutex Global\HuntressBR13_000, url https://node-000.day13.example.invalid/api/BR13-000, CRC-32 0xa69b7e61;
• HCFG record at file offset 560,188 (Day 12's Pascal build had it at 82,928); SHA-256 6d141b21…e0df.

Unconfirmed: the Crystal runtime details the prompt hints at (symbols, sections, file size). No question asked about them, and the file wasn't reopened for this write-up.

The timed variants used the same template with different values. The Time Trial's Day 13 file was variant 007 (Global\HuntressBR13_007, 0x6c8999d9), still at offset 560,188, and so was the Mastery's set 3.

Single Solve (the solver, 3 batches)

The core already covered all 10 new fields. No new add-on was written. The morning solver picked up the new challenge by itself with the core Binary solver, sent the core's answers first, then the two add-ons'. Every batch was accepted:

The d01 ELF batch included ELF64, once rejected on a different question; it was accepted again. Nothing was rejected. CORRECT at 13:00:47.698 UTC, flag 0f28af3e…00b0.

Time Trial

The start reply said unique_attempts_remaining: 0 and unranked. The clock was 80 s (Day 12's Binary routes had 30 s).

started_at   2026-10-13T13:00:54.283Z   deadline 13:02:14.283Z (80 s)
artifact     evidence.zip, 32.2 MB, Day 13 = variant 007
answers      130, one submit
             d13_analysis_config_mutex: Global\HuntressBR13_007
             d13_analysis_config_url: https://node-007.day13.example.invalid/api/BR13-007
             d13_analysis_config_crc32_hex: 0x6c8999d9
             d13_analysis_hcfg_offset: 560188
laps         start 463, artifact 833, solve 599, ready 0, submit 840, TOTAL 2735 ms
vs started_at (client clock): start +115, artifact +949, solve +1547, ready +1548, submit +2387 ms
result       CORRECT 13:00:56.091Z   flag 78198067…483d

Mastery

started_at   2026-10-13T13:01:03.071Z   deadline 13:04:23.071Z (200 s)
artifacts    3 streamed downloads: 01/02/03-evidence.zip, 32.2 / 32.3 / 32.2 MB
answers      390 (130 per set), one submit; e.g. set3.d13_analysis_hcfg_offset: 560188
laps         start 894, dl3 801, dl2 692, dl1 393, artifact 1, solve 538, ready 1, submit 1329, TOTAL 4648 ms
vs started_at (client clock): start +235, dl3 +1036, dl2 +1728, dl1 +2121, artifact +2122, solve +2660,
             ready +2661, submit +3989 ms
result       CORRECT 13:01:06.370Z   flag cc96598c…c14b

The three downloads ran at the same time and finished out of order (3, 2, then 1). About half the clock went to downloads (2.1 s of 4.0 s).

Later: ranked retries

On 2026-10-15 Days 13–14 began to accept ranked-clock requests with fresh pools. The Mastery now came as five zips (650 answers, 320 s clock). On a retry, the submit reply's solved_by.at still shows the first CORRECT, so these are our own client-side laps.

Boards after all of 2026-10-15's passes: Time Trial 4th of 15 (2.646 s), Mastery 2nd of 11 (4.302 s). The top times (0.155 s and 0.224 s) are too fast to include a full download of a multi-MB zip; we haven't worked out how.

Time to solve

The morning solver started the Single Solve at 13:00:42 UTC and had it CORRECT 5.3 s later; the Time Trial (1.8 s) and Mastery (3.3 s) each went CORRECT on the first submit, unranked, for 23.9 s from the first start to the last CORRECT. Ranked retries on 2026-10-15 ET (three passes, all CORRECT) set the board times above.

Takeaways

• The prompt's language doesn't decide the questions. "Fingerprint the runtime" was never asked; the questions were the usual field set, and the parser that read Day 12's Pascal build read Day 13's Crystal build unchanged.
• The same record, a bigger program. The HCFG record moved from offset 82,928 to 560,188; a search by structure and CRC doesn't care where it sits.
• Mastery is download-bound. Solve was ~0.5 s; downloads were half the day's run and most of the slow ranked one.