session impression: a hinted session and a separate case portal

huntress ctf 2026, day 21 · web exploitation · all three modes solved

tools used: live web target capture.

the main target supplied a session hint that opened a protected secondary route. mastery was deliberately heterogeneous: it could also assign a backup site or a case-management application with a different authorization flaw.

the key insight was to treat the clue-driven session flow and the case portal as distinct applications, while keeping each target's cookies and evidence separate.

challenge shape

the single and time trial used the hinted-session application. mastery could bundle that pattern with an uptime backup and a case portal, so the challenge name alone was not a sufficient classifier.

artifact anatomy

the session clues supplied a feature route, a secondary route, and a hint. the feature route accepted the hint as a query parameter and set an authenticated cookie for the target. the mastery case portal instead established its own cookie on the landing page, then exposed case metadata through an event feed.

extraction method

for the session target, request the feature route with the clue-provided hint and reuse the same target-local cookie jar on the protected route. for the case portal, read the authenticated activity feed, prioritize reference-only or other-owner records, and request those incident ids. a backup target follows its own robots.txt entry and indexed backup file.

get(featureRoute, params={"sid": sessionHint})
proof = get(secondaryRoute)  # same target client retains its cookie

events = get_json("/api/events")
for event in prioritize_reference_only(events):
    record = get_json(f"/api/incidents/{event['case_id']}")

validation and answer formatting

each answer comes from the current target's protected route, incident record, or backup response. the accepted results are [redacted]; no target proof is copied into this post.

failed approaches and pivots

no request transcript or failed submission was preserved in this checkout, so there is no challenge-specific failed pivot that can be verified. the two session/case flows are documented from the recovered solver description and covered by local target fixtures.

automation across modes

the collector dispatches from clue keys when available, then uses target-local page evidence and the known challenge family only for sparse targets. every mastery target has an independent client and cookie jar.

lessons

solve times

public scoreboard · day 21 finish: below top 10

  • 23:04.0single#61 of 163 solvers
  • 24:39.8time trial#61 of 160 solvers
  • 27:06.5mastery#59 of 157 solvers
  • #36team day finishof 146 teams that cleared it, in 27:16.2
  • #8single orderof the day's 8, by our flag time

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

work-to-flag, attempts, and wrong submits were not recorded in this checkout.

mode work to flag attempts wrong submits
single not recorded not recorded not recorded
time trial not recorded not recorded not recorded
mastery not recorded not recorded not recorded

the clock

the winning attempt start, deadline, and flag timestamp were not preserved, so available, used, and remaining time cannot be reported.

mode time available time used time left
time trial not recorded not recorded not recorded
mastery not recorded not recorded not recorded

where we got stuck

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

{generated by a clanker}