🌐 If a Breach Falls in a Log — Day 25 Web Exploitation
- Challenge ID:
web-exploitation-day-25-if-a-breach-falls-in-a-log - Lane:
web-exploitation - Station: Homestation (HS Web lane)
- Solved by: HS (All three modes CORRECT per GM)
- Writeup salvage source: HS Homebase VM at
[local path](preserved into the repo 2026-10-28).
Challenge prompt
"You broke in, and nobody wrote it down. Find the decisive moment the logs missed — if a breach falls in the forest, did it make a sound?"
No downloadable artifact; each attempt assigns live targets. SS and TT: one question (proof-1). Mastery: five targets (proof-1…proof-5).
Results
| Mode | Result | Game Master flag |
|---|---|---|
| Single Solve | CORRECT | Token64fd1d5ae5f958fa2fc94fc2164b1c0f |
| Time Trial | CORRECT | Token19f979673a341ee82a4a8bc5b5d734bb |
| Mastery | CORRECT | Token9e6d4bb5197c321ce0ca424ca14e2b09 |
Time Trial defaulted to run_for_time: false (no deadline). Mastery also had no deadline.
Method summary
Single Solve + Time Trial (same audit flow)
- GET
/api/cluesto obtain the feature route + secondary audit route. - POST
{}to the feature route → response reports a completedprotected_exportand anexport_id. - GET the secondary (audit) route → find the event for that export (
page_viewin these targets). - POST the secondary route with:
{
"export_id": "<ID from export response>",
"missing_event": "protected_export",
"observed_event": "page_view"
}
- Submit the returned
evidenceasproof-1.
Mastery (5 targets, 5 different web flaws)
| Answer | Technique | Proof |
|---|---|---|
proof-1 | robots.txt → /archive/ dir listing → JSON .bak file leaks the proof | Proofproof_579e1a0bebbe4e3d83717dcda17acc53 |
proof-2 | Home-page case_session cookie → /api/events reveals case ref → /api/incidents/<id> returns the restricted case | Proofproof_5d1dde45bee7424ea25ba9657bb52763 |
proof-3 | /explore?q= SQL injection: ') UNION SELECT id,title,topic,summary FROM articles-- | Proofproof_50a0c974570c4b66bf51721ec8549937 |
proof-4 | /notes.php?doc=../../../../run/huntress/objective.txt LFI | Proofproof_650956772ca14f949556b5f64c7c265f |
proof-5 | Repeat the audit flow (feature + secondary route) against Mastery target 5 | Proofproof_888b95b3e55146459c4116fcb4148812 |
SS proof: Proofproof_9c4756c61c784505a5979b82c5b0c3ac TT proof: Proofproof_006d788a634e4913a6a5c623433d528b
The UNION query's closing parenthesis was mandatory for this target's query context. sqlite_master enumeration identified the articles table with columns id, title, topic, summary, visibility.
Lessons / faster-next-time
- Start each mode first and use its assigned target URLs — these change per attempt.
- Read the target's visible route inventory (
/api/clues,robots.txt, home-page links) BEFORE guessing paths. - For the audit flow: copy the LIVE
export_id+ observedeventfrom the response into the report request. Values are per-target, not reusable. - Session-based APIs: GET
/first with a cookie jar, then hit the API. - SQL injection: confirm the schema via
sqlite_master, THEN UNION with the right column count/paren context. - Mastery can be a MIX of vuln types — do not assume one exploit applies to all 5 targets.
Script
scripts/solve.py — dependency-free Python 3 proof collector. Supports:
python3 solve.py audit TARGET_URL # audit-flow (SS/TT and MM target 5)
python3 solve.py archive TARGET_URL # robots.txt + archive listing + .bak
python3 solve.py idor TARGET_URL # case_session + /api/events + /api/incidents
python3 solve.py sqli TARGET_URL # UNION injection
python3 solve.py lfi TARGET_URL # ?doc= traversal
python3 solve.py mastery U1 U2 U3 U4 U5 # all 5 targets at once
The solver prints proofs; it does not submit to GM. autosolve.py submits.
HS lane adoption
Same autosolve_web.py extraction logic as Fail Open Friday — these 5 patterns (archive leak, IDOR, SQLi, LFI, audit-event flow) recur. Any autosolve_web.py that learns BOTH days' probe shapes covers both; TT and MM are per-attempt mutations with the same route→flaw mapping.