2026-10-07
Day 7: Routine Check
Lane: Binary reverse engineering. Modes: Single Solve, Time Trial, Mastery. Result: all CORRECT.
The challenge
- ID:
binary-reverse-engineering-day-07-kotlin-staging-logic - Prompt: "Kotlin coroutines don't run — they suspend, resume, and hop between states. Reconstruct the state machine from the metadata and figure out what this one was staging."
What we were given
- One
evidence.zip(6,485,485 bytes, SHA-256a9cd8efd7266ea8dc6972ae9112dbfaaa3f90c2f6129c6013d124c71f5d1f450). AES encrypted, passwordinfected. - The earlier files again:
d01(C),d02(C++),d03(.NET),d04(Go),d05(Rust),d06(Java). - New today:
d07--variant-000.jar(5,152,890 bytes), a Kotlin program packed as a JAR. - It bundles the Kotlin runtime: 3,244
kotlin/entries, includingkotlin/coroutines/Continuation.class. - The program classes are
Day07Variant000.classplus the coroutine classesDay07Variant000$main$1,$main$2and$staged$1.
What was asked
- 56 questions in Single Solve and Time Trial. Mastery had two deliveries, so 112.
- Question types:
- Hashes: SHA-256 of each member.
- d01 to d06: the same header, structure and config questions as before.
- d07 (new): the archive format, the config fields
campaign,domain,mutex,url, the CRC32, and theHCFGversion byte.
How it was solved, start to finish
- Ran the driver
run_tm.py: start the attempt, download the zip, runsolve_tm.py, submit. - Routine fields for d01–d06 were answered as before.
- d07 (the new part): the Kotlin JAR hides its config the same way as the Day 6 Java JAR. The question gave the same recipe:
- Open the
Main-Classclass file and walk its constant pool. - Base64-decode each UTF-8 constant of 80+ Base64 characters.
- Accept the first result that is a CRC32-valid
HCFGrecord, then XOR-decode and parse the JSON.
The Day 6 code path handles it. Re-checked on the saved sample:
[jar:Day07Variant000.class base64 run at file offset 206, run length 508]
HCFG version=1 xor_key=0x31 len=368 crc32=0xcf3ea232
JSON: {"campaign":"BR07-000","domain":"node-000.day07.example.invalid","ipv4":"198.51.100.140","url":"https://node-000.day07.example.invalid/api/BR07-000","mutex":"Global\\HuntressBR07_000", ...}
Answers: format Java archive, CRC32 0xcf3ea232, HCFG version 1, campaign BR07-000.
- Submitted all 56 answers (
answers.json). Result:CORRECT. - Time Trial: a fresh variant, same steps. Result:
CORRECT. - Mastery: two deliveries (112 answers), same steps. Result:
CORRECT.
Results
| Mode | Result | Flag |
|---|---|---|
| Single Solve | CORRECT | c0813d767b25aadd8074e4cdfefcc5b7 |
| Time Trial | CORRECT | fa77af724f2534548b9785781f5f1de0 |
| Mastery | CORRECT | 8a588fb8b39eb0072baabcc20929759d |
Each mode folder has the raw start-response.json, answers.json, submit-response.json and final-status-response.json.
Notes on the saved files:
mastery/answers.jsonwas saved empty.mastery/final-status-response.jsonshows all 112 answers accepted.mastery/start-response.jsonis from a later restart that the Game Master refused (rotation_exhausted).
Reproduce
Offline, from this folder (no Game Master needed):
mkdir -p /tmp/d07/deliveries && cp single-solve/evidence.zip /tmp/d07/deliveries/delivery-1.zip
python3 -I ~/.config/huntress/solvers/solve_tm.py single-solve/start-response.json /tmp/d07/deliveries /tmp/d07/answers.json
Checked on 2026-11-01: this gives the same 56 answers as answers.json. It only parses the samples and never runs them.
Live driver:
python3 -I ~/.config/huntress/solvers/run_tm.py binary-reverse-engineering-day-07-kotlin-staging-logic single_solve /tmp/d07
This needs a Game Master token in ~/.config/huntress/gm-token.
Day 7: Seven-Zip, No Relation
Lane: Metadata forensics. Modes: Single Solve, Time Trial, Mastery. Result: all CORRECT.
The challenge
- ID:
metadata-forensics-day-07-seven-zip-no-relation - Prompt: "Files arrive inside and beside a collection of ZIP archives. Count what the archives claim, distinguish empty space from evidence, and find the census that does not add up."
What we were given
- Single Solve:
(evidence.zipmetadata-day-07-evidence.zip, 339 bytes). It holds one file:evidence-drop-001.zip. - Time Trial:
metadata-day-07-evidence.zip(3,309 bytes, SHA-256dbfe9dcd1c89e310ef9812bcde1bb99515101de45da384841ff5758692c32b00). It holds five files:evidence-note.txt,camera-001.jpg,panel-001.png,document-001.pdfandevidence-drop-001.zip. - Mastery:
metadata-forensics-day-07-seven-zip-no-relation-mastery.zip(6,916 bytes, SHA-256e6790598cee344af5566559c4fd0e7795617adf08398551d696ad860fce2b00b). It holds two inner zips,metadata-day-07-evidence-01.zipand-02.zip, each with the same five file names. - New today: a zip inside the zip.
What was asked
- Single Solve: 1 question:
zip.eocd_total_entriesforevidence-drop-001.zip. - Time Trial: 11 questions. This is the first "composite" round: one question for every field type seen so far.
- Mastery: 22 questions, 11 per set (
set1.,set2.). - Question types:
- File identity (evidence note): filename, size, MD5, SHA-1, SHA-256, first 16 bytes as hex, UTF-8 BOM present.
- Headers: JPEG SOF size, PNG IHDR size, PDF header version.
- New: ZIP End Of Central Directory (EOCD) total entries.
How it was solved, start to finish
- Single Solve, by hand. Extracted
evidence-drop-001.zipand read its EOCD record. The EOCD starts withPK\x05\x06. Search from the end, then read the 2-byte "total entries" field at offset +10:
i = b.rfind(b'PK\x05\x06')
total = struct.unpack_from('<H', b, i + 10)[0]
Result: 2. Submitted as d07-zip-001-zip-eocd-total-entries. CORRECT.
- First try at the timed modes. Added a
zip_eocdsolver to the one-answer driverrun_mode.py. It returned no result: the Time Trial asked 11 different questions about 5 files, not one. - Wrote
run_meta.py, a composite driver. It does the whole loop for any mix of fields: - Start the attempt (
op=startfor Single Solve,op=retryfor the timed modes). - Download the bundle.
- For each question, take the file name after
->in itsartifactfield (for examplemetadata-day-07-evidence.zip -> camera-001.jpg). Find it withfind_file, which extracts with7z x -y -pinfectedand searches nested zips. - Pick a handler from the answer ID. For example
...-md5→hashlib.md5,...-magic-prefix-hex-16→fb[:16].hex(),...sof-dimensions...→ JPEG SOF parse,...eocd-total-entries→ the EOCD read above. - Submit every answer in one JSON object.
- Time Trial.
run_meta.pyanswered all 11. Examples fromanswers.json:
evidence-note-size-bytes=33evidence-note-magic-prefix-hex-16=43617365204333413635314546324441d04-jpeg-001-jpeg-sof-dimensions-pixels=257x146d07-zip-001-zip-eocd-total-entries=2
CORRECT.
- Mastery. IDs start with
set1.orset2.. Each set has its own inner zip. The driver maps set N to the N-th inner zip, strips the prefix, computes the answer, and puts the prefix back. All 22 answered.CORRECT.
Results
| Mode | Result | Flag |
|---|---|---|
| Single Solve | CORRECT | a402f1dbcfd5ccc361b3e2157b46d39f |
| Time Trial | CORRECT | cc3ed4df69afb1d06de32b2a56f9e589 |
| Mastery | CORRECT | f0f2e213eb7669f064fe7fcf8ce5b1d9 |
Each mode folder has answers.json, start-response.json, submit-response.json and final-status-response.json. The Single Solve folder also has the zip; its saved submit says already_solved because it is a re-run, and the final status shows CORRECT. No screenshots were captured.
Reproduce
python3 -I ~/.config/huntress/solvers/run_meta.py metadata-forensics-day-07-seven-zip-no-relation time_trial /tmp/d07
This needs a Game Master token in ~/.config/huntress/gm-token. For time_trial and mastery it calls op=retry first, which spends one of the unique attempts. It only parses the files. It never runs them.
Day 7: WebView Bridge
Lane: Mobile (Android). Modes: Single Solve, Time Trial, Mastery. Result: all CORRECT.
The challenge
- ID:
mobile-android-day-07-webview-bridge - Prompt: "Cross a challenge-owned WebView bridge through an untrusted origin."
What we were given
- One Android APK per case. Inside:
AndroidManifest.xml,classes.dex,assets/huntress_puzzle.json, andMETA-INF/signing files. - Single Solve:
webview_bridge.apk, caseMOB-07-9F5C796C19(12,704 bytes, SHA-256160ca9fa498ffbd877275ba226dde6f919d1a20511c3c1ed6b00811ca44df55a). Its puzzle is saved assingle-solve/huntress_puzzle.json. - Time Trial: one APK (case
MOB-07-4800A56C6D). Mastery: three APKs. These APKs were not saved in the repo.
What was asked
- One question per case: return the exact backend answer code for that case.
- Answer format:
ANS-plus exactly 16 uppercase hex characters. - Answer IDs:
answerfor Single Solve and Time Trial.set1.answertoset3.answerfor Mastery. - The saved Time Trial and Mastery attempts had no deadline (
run_for_time: false).
How it was solved, start to finish
The app was never installed or run.
- Started the Single Solve and downloaded the APK.
- Read the puzzle asset. New codec today:
caesar.
{"schema":"huntress.mobile.puzzle.v1","day":7,"codec":"caesar","data":"KX]7BKPOK?>:CCP=;:@O","key":10}
- Decoded it. Shift every character back by
key(10). For exampleK(75) becomesA(65).
import json
p = json.load(open('single-solve/huntress_puzzle.json'))
print(''.join(chr((ord(c) - p['key']) % 256) for c in p['data']))
Output: ANS-8AFEA54099F3106E.
- Submitted it under
answer. The Game Master repliedCORRECT. - Taught the driver. Added the
caesarcodec tomobile_puzzle()inrun_mode.py. - Time Trial and Mastery: ran
run_mode.pyon both. Both came backCORRECT. The Time Trial answer wasANS-EF4F4D6A583828D4.
Results
| Mode | Result | Flag |
|---|---|---|
| Single Solve | CORRECT | 6640e170fc7ef9afdc8e733b3d022570 |
| Time Trial | CORRECT | 89e1a6e5b7a1eee5ffae6b49aae0a7f7 |
| Mastery | CORRECT | 71642323484b46ddfb868a5acd814c6f |
Each mode folder has answers.json, start-response.json, submit-response.json and final-status-response.json. The Single Solve submit-response.json says already_solved because it is a resubmit. Its final-status-response.json holds the CORRECT result and flag. No screenshots were taken.
Reproduce
Offline: run the Python snippet above from this folder.
A fresh run with the driver:
python3 -I ~/.config/huntress/solvers/run_mode.py mobile-android-day-07-webview-bridge single_solve mobile_puzzle /tmp/d07
This needs a Game Master token in ~/.config/huntress/gm-token. It only unzips the APK and reads the asset. It never installs or runs the app.
Day 7: Node by Node
Lane: Script analysis. Modes: Single Solve, Time Trial, Mastery. Result: all CORRECT.
The challenge
- ID:
script-analysis-day-07-node-operation-graphs - Prompt: "Eight Node.js operation graphs that will never execute. Spot the indirect expansions and token fragmentation, then sort each graph into its role — benign, staging, dropping, downloading, uploading, launching, scheduled, or multi-stage — from the final bytes alone."
What we were given
- One
evidence.zip(66,464 bytes, SHA-2565e2cadf1264bbf1869647fab33b9f630d7ace1d513d4a21c0f7e400040c61038). Password:infected. - The Day 6 set, plus two kinds of new files:
- 8 Node.js files (
node-*.mjs). SameMaterial_/Codec_pattern as Days 4–6, in double quotes. New material formats include\xNNescapes (escaped_octets) and JSON hex arrays (hex_array). Real codec lines:
escaped_octets|rolling_xor|ec
hex_array|single_byte_xor|3b
hex_buffer|reverse_bytes|
base64_chunks|reverse_bytes|
- 2 "advanced" Bash files:
indirect-expansion.shandtoken-fragmentation.sh.
What was asked
- 469 questions: the 399 from Day 6, 56 Node.js questions (7 per file) and 14 advanced-Bash questions.
- Node.js questions per file: role, ATT&CK ID, adapter, campaign, steps (joined by
>), stage count, plus one indicator (URL, file path, host ID or schedule path). - Advanced Bash questions per file: role, ATT&CK ID, campaign, steps, stage count, plus the feature and the primitive used.
How it was solved, start to finish
- Started the untimed Single Solve and downloaded the archive.
- Wrote a Node.js decoder (
nodedec.py). It is the Bash decoder with extra material parsers:
\uNNNN(unicode_escapes),\xNN(escaped_octets), JSON hex arrays (hex_array), and%NNstrings.- Three files failed on the first pass because of the two new formats. Once those parsers were added, all 8 decoded.
- New role: persistence.
node-schedule-descriptor.mjsdecodes tobuild_schedule,register_descriptor. That ispersistence_descriptor, mapped toT1053.003(Scheduled Task/Job: Cron). Node's interpreter ID isT1059.007.
| File | Steps | Role | ATT&CK |
|---|---|---|---|
node-context.mjs | decode_configuration>return_configuration | benign_utility | none |
node-memory-stage.mjs | acquire_remote>dispatch_memory | stager | T1105 |
node-embedded-write.mjs | decode_embedded>write_file | dropper | T1140 |
node-remote-write.mjs | acquire_remote>write_file | downloader | T1105 |
node-outbound-transfer.mjs | read_file>transmit_remote | uploader | T1041 |
node-local-dispatch.mjs | locate_local>dispatch_process | launcher | T1059.007 |
node-schedule-descriptor.mjs | build_schedule>register_descriptor | persistence_descriptor | T1053.003 |
node-staged-chain.mjs | profile_host>acquire_remote>write_file>change_mode>dispatch_process | staged_chain | T1059.007 |
- Advanced Bash. These configs start with
v=1|feature=..., notv=1|steps=.... The old check rejected them, so all decoders were changed to accept anyv=<n>|<key>=start. Decoded:
indirect-expansion.sh: v=1|feature=indirect_expansion|primitive=parameter_indirection|campaign_id=BASHX-F434AEA9|stage_count=1|steps=recognize_syntax,reconstruct_intent
token-fragmentation.sh: v=1|feature=token_fragmentation|primitive=array_token_construction|campaign_id=BASHX-E4F52183|stage_count=1|steps=recognize_syntax,reconstruct_intent
Their role is the special value advanced_syntax, and ATT&CK is none.
- Two wiring fixes:
- Node answer IDs end in multi-word fields (
host-id,file-path,stage-count), so the suffix matching was fixed. - The
nodejs-prefix was added to the dispatch loop.
After that, all 469 answers were CORRECT. Real answers:
{"nodejs-persistence-role": "persistence_descriptor", "nodejs-persistence-attack": "T1053.003",
"nodejs-persistence-adapter": "systemd_descriptor", "nodejs-persistence-schedule-path": "/etc/cron.d/node-f4b470",
"bash-advanced-indirect-role": "advanced_syntax", "bash-advanced-indirect-feature": "indirect_expansion",
"bash-advanced-indirect-primitive": "parameter_indirection"}
- Time Trial: five answers were missing at first, because the Node chain file did not decode in that variant. After that fix,
token-fragmentation.shfailed on the same variant. The material parser was made more general, and the Time Trial passed. - Mastery (938 questions, two sets):
- The first try hit a 409 because another attempt was still active.
- Then set 2's
indirect-expansion.shused a new format,reversed_base64: the Base64 text itself is reversed before decoding. Reversing bytes afterwards is not the same thing, so the Base64 branch of every decoder was patched. - Mastery then passed.
- The saved Time Trial and Mastery attempts for this day have no deadline (
run_for_time: false).
Results
| Mode | Result | Flag |
|---|---|---|
| Single Solve | CORRECT | d6506462aeb2a3e5c45b9f35274c81d1 |
| Time Trial | CORRECT | 6b39c0493fdaf4974978f1b7cb6654f3 |
| Mastery | CORRECT | 0a3cf17de369caad0674de448e827843 |
Each mode folder has answers.json and the Game Master JSON. The Single Solve folder also keeps evidence.zip. No screenshots were taken.
Reproduce
Offline, from the repo root:
mkdir -p /tmp/d07/ex
7z x -pinfected -o/tmp/d07/ex repos/day-07-script-analysis-node-by-node/single-solve/evidence.zip
python3 -I tools/solve_expansion.py /tmp/d07/ex \
repos/day-07-script-analysis-node-by-node/single-solve/start-response.json /tmp/d07/answers.json
This rebuilds all 469 saved answers (re-checked on 2026-11-01). It only parses the files.
Live: python3 -I ~/.config/huntress/solvers/run_expansion.py script-analysis-day-07-node-operation-graphs <mode> <workdir>. This needs a Game Master token. The solvers/ copy runs JScript fixtures in a Node vm sandbox; the tools/ copy does not.
Day 7: Percent Chance
Lane: Warmups. Modes: Single Solve, Time Trial, Mastery. Result: all CORRECT.
The challenge
- ID:
warmups-day-07 - Prompt: "Half the web runs on this encoding. One percent at a time."
- New encoding: URL percent-encoding
What we were given
- One blob file per mode (Single Solve and Time Trial), or a zip of multiple blobs (Mastery).
- Each blob has numbered lines (
NN<TAB>fragment). Each line is independently encoded. - The decoded text ends with a slug like
word-word-word-NNto submit.
How it was solved
- Decode the blob. Standard percent-encoding (
%XXhex pairs). URL-decode to get plaintext. - Mastery blobs use multiple encodings from Days 1-7. The solver (
warmup_decode.py) tries every known decoder on each line and picks the one producing printable text. - Concatenate all decoded lines. The text contains a message ending in the answer slug.
- Submit the slug to the Game Master.
Results
- Single Solve:
CORRECT. slug:witty-storm-fern-77. Flag:3733b5898515b564508907ceb4f0c6f5. - Time Trial:
CORRECT. slug:happy-starfish-quartz-38. Flag:c24b1d8cd18a850e5bba6efaeca98f14. - Mastery:
CORRECT. 33 slugs. Flag:512822f1b436ef3ef4d8c46a21841441.