next, next, finish: an nsis script with something extra

huntress ctf 2026, day 22 · script analysis · all three modes

tools used: script analysis and timed-attempt pipeline.

"an nsis installer script clicks through its own setup wizard while carrying a little something extra. separate the real setup actions from the modeled payload behavior, and recover every cross-language handoff from the final bytes." the day's new carrier was the nullsoft installer's .nsi script, beside every earlier day's samples.

the key insight: a new carrier language is a new comment and string syntax around the same analysis header, not a new problem. the descriptors the questions ask for sit in that header in every carrier.

challenge shape

single        1417 fields   every earlier sample's descriptors plus the
                            day's polyglot-nsis-<role>-<field> family
time trial    1417 fields   the same over a fresh bundle
mastery       8502 fields   six bundles

44 fields were new: four nsis specimens (benign utility, downloader, launcher, staged chain) at eleven fields each.

artifact anatomy

the bundle holds every earlier specimen and four nsis scripts, one per role. nsis takes ; and # line comments and /* */ blocks; these scripts are seven lines and run nothing:

; HUNTRESS_ANALYSIS_V1:<CODEC>-<payload>
Unicode True
Name "Huntress Analysis <hex>"
OutFile "huntress-<hex>.exe"
Section
  StrCpy $0 "<the same payload>"
SectionEnd

each script used a different codec (B64, HEX, HB64, FRAG). the "handoff" is the descriptor's next_language and stage_profile, not a command in the script. the fields ask the role, the attack technique, the adapter, the campaign, one indicator (a file path for the benign utility, a url for the rest), the stage count, the steps, the host profile, the obfuscation, the stage profile and the next language.

extraction method

the script analysis reader finds the HUNTRESS_ANALYSIS_V1:<CODEC>-<payload> header by its marker whatever comment character precedes it, decodes the payload with the codec it names, and answers each descriptor key by name, the next language included. the reader needed no change for the new carrier.

validation and answer formatting

labelled fields with an enum in the question are answered from that list only; free-text fields (urls, paths) are answered as stored.

failed approaches and pivots

none on day 22.

automation across modes

the same reader answers all 1417 fields of every bundle.

lessons

solve times

public scoreboard · day 22 finish: below top 10

  • 0:04.9single#3 of 163 solvers
  • 0:07.4time trialfirst blood, +0.58 s
  • 0:12.2mastery#3 of 158 solvers
  • #26team day finishof 145 teams that cleared it, in 10:39.5
  • #4single orderof the day's 7, by our flag time

from the public scoreboard: 09:00 et release to our flag; the table below counts from our attempt's start.

from when we started working on the mode to its flag.

mode work to flag attempts wrong submits
single 2.2s 1 0
time trial 1.1s 1 0
mastery 2.8s 1 0

the clock

server timer of the winning attempt; used is attempt start to flag, left is what remained on the clock.

mode time available time used time left
time trial 95.0s 1.1s 93.9s
mastery 200.0s 2.8s 197.2s

where we got stuck

flags: single, time trial, and mastery [redacted].

{generated by a clanker}