ctf-methodology
CTF challenge triage and time management — identify the category and intended technique from the handout artifacts in the first 10 minutes, budget points vs. time, and know when to park a challenge. L
SKILL.mdskills/ctf/ctf-methodology/SKILL.md
GitHubCTF methodology: triage and time budgeting
When it applies
You're working jeopardy-style CTF challenges (authorized by definition — the organizer hands you the
targets). This skill is the entry point: it classifies the challenge, picks the category playbook
(ctf-web, ctf-pwn, ctf-crypto, ctf-rev, ctf-forensics), and keeps you from sinking three
hours into a 100-point guessy challenge.
Why it works
CTF challenges are designed puzzles with an intended path, and the handout almost always telegraphs it: the category, the files provided, the service type, the description's wordplay. Reading those signals first converts "stare at an unknown blob" into a 2-minute routing decision, and a point/time budget converts sunk-cost grinding into deliberate moves.
Method
- Read the handout like a spec, not a flavor text. Category, title, and description are hints:
puns name the technique ("relationships" → RSA common modulus, "padding" → padding oracle),
an unusual port or protocol is the challenge. Note the flag format (
flag{,CTF{) so you can grep for it later. - Inventory the artifacts (2 minutes). For each provided file:
Archive/disk image →file <f> ; exiftool <f> ; strings -n 6 <f> | head -50 ; checksec --file=<bin>ctf-forensics. ELF/PE/APK with a flag-check →ctf-rev. ELF +nc+ libc →ctf-pwn..py/.sagewith math →ctf-crypto. URL only →ctf-web. - Fingerprint the service.
ncbanner grab; for HTTP,curl -sS -D- <url>and read headers, cookies, comments, and linked assets — server header (Flask/PHP/Express) narrows the bug class immediately (recon-techstack-fingerprinting). - Classify difficulty honestly. Provided source + small binary + common category = intended easy; read the source for the bug before touching tools. No source + large binary + kernel/heap/ custom crypto = intended hard; expect the full chain and budget accordingly.
- Budget by points and momentum. A working heuristic: cap exploration at ~15–20 minutes per untried idea, and re-evaluate a challenge when you've spent more time than its points justify relative to what remains open. Points are worth the same everywhere — three quick 100s beat one stalled 400.
- Move on deliberately, not silently. Park a challenge when: you've enumerated the obvious surface twice with no new information, every hypothesis needs information the challenge doesn't provide (guessy), or a teammate/other challenge offers better points-per-hour. Write down the state (what you tried, what's ruled out, next idea) so re-entry is cheap — solves often come from a fresh look or a later organizer hint.
- Watch for released hints and solve counts. Organizers drop hints when a challenge stalls event-wide; a high solve count means the intended path is simple and you're overthinking — go back to the description and the most obvious reading.
Gotchas
- The category label lies occasionally — "web" challenges that end in SSTI-to-RCE, "crypto" that's really encoding, "misc" that's actually forensics. Re-classify as evidence appears instead of forcing the labeled playbook.
- Warmup/100-point challenges never require deep tooling. If you're writing a heap exploit for a warmup, you missed something — look for the joke, the encoding, the default creds.
- Guessy challenges (stego with no cover text, obscure cipher with no key hint) burn unlimited time. Timebox hard and return after everything tractable is done.
- Dynamic/shared instances — other players can leave artifacts in shared services (uploaded shells, changed passwords); a freshly spawned instance is worth the restart.
- Don't skip the flag format grep —
strings, pcap searches, and memory dumps often yield the flag directly togrep -a 'flag{'before any "real" solve.
Verify success
Every open challenge has: a category assignment, a loaded playbook, a written list of tried/ruled- out ideas, and an explicit keep/park decision — no challenge is being worked by vibes. Scoreboard confirms submitted flags.
References
Routes into ctf-web, ctf-pwn, ctf-crypto, ctf-rev, ctf-forensics; technique depth in the
domain skills they chain to (web-testing-checklist, exploit-pwn-chain, crypto-rsa-attacks,
reverse-eng-binary-triage, defense-dfir-triage).