본문으로 건너뛰기
cccskills
무료GitHub에 공개

lookup-ransomwarelive

Use when you need to check whether an organisation/domain has been claimed by a ransomware group, profile a specific ransomware group (TTPs, leak-site infra, IOCs, YARA), or aggregate ransomware victim claims by country/sector/timeframe. Backed by ransomware.live's leak-site scrapes — 27k+ victims across 330+ groups. Commonly invoked by /domain-investigation, /ransomware-ecosystem, /threat-actor-profiling, and detection-engineering workflows. Reads $RANSOMWARE_LIVE.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

실행 점수 미검증

설치 방법 보기

포함된 파일(1)

  • SKILL.md8.0 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

lookup-ransomwarelive

Queries ransomware.live's PRO API for victim claims, group profiles, IOCs, YARA rules, ransom notes, negotiation chats, and CSIRT contacts. Retrieval only — the invoking skill or agent reasons about the result. Note that leak-site claims are claims, not confirmed breaches (see Source reliability below).

When to invoke

  • An investigated domain or organisation should be checked against ransomware leak-site claims
  • Profiling a ransomware group — pull description, TTPs, tools, vulnerabilities, leak-site .onion infrastructure
  • Sector- or country-level ransomware briefing (e.g., "all NL victims claimed in 2026")
  • Detection engineering for a specific group — fetch its public YARA rules + IOC dump
  • Updating the /ransomware-ecosystem knowledge cell with the current top-N groups by victim count
  • Incident response: looking up a country's CSIRT/CERT contact list

Do NOT invoke for:

  • General malware classification — use /lookup-virustotal
  • Confirming whether a specific incident actually occurred — leak-site claims are unverified by default
  • Initial-access vector intelligence — that's not on leak sites; chain /initial-access-brokers instead

How to invoke

Single Python CLI (stdlib only — no install).

Sanity checks

python3 tools/clis/ransomwarelive.py validate     # check the API key
python3 tools/clis/ransomwarelive.py stats        # global totals

Victim lookups

# Free-text search across victim names + descriptions
python3 tools/clis/ransomwarelive.py search --q "acme corp"

# Targeted filters (combine freely)
python3 tools/clis/ransomwarelive.py search --country NL --limit 10
python3 tools/clis/ransomwarelive.py search --sector Healthcare --country US
python3 tools/clis/ransomwarelive.py search --group lockbit3 --limit 5

# Last 100 claims globally (newest first)
python3 tools/clis/ransomwarelive.py recent --limit 20

# Single victim detail by ransomware.live id
python3 tools/clis/ransomwarelive.py victim "QksgR3JvdXBAYWtpcmE="

Group intelligence

# All 330+ groups, sorted by victim count
python3 tools/clis/ransomwarelive.py groups --limit 20

# Profile of a specific group — description, TTPs, tools, leak-site URLs, etc.
python3 tools/clis/ransomwarelive.py group-profile lockbit3

# Lighter group detail
python3 tools/clis/ransomwarelive.py group lockbit3

Defensive intel

# IOC dump (md5, ip, etc.) — all groups
python3 tools/clis/ransomwarelive.py iocs
# Per-group IOCs
python3 tools/clis/ransomwarelive.py iocs lockbit3

# YARA rules — all groups, or per group
python3 tools/clis/ransomwarelive.py yara
python3 tools/clis/ransomwarelive.py yara lockbit3        # full rule content

# Ransom note samples
python3 tools/clis/ransomwarelive.py ransomnotes lockbit3

# Negotiation chat logs (if available for that group)
python3 tools/clis/ransomwarelive.py negotiations lockbit3

Context

python3 tools/clis/ransomwarelive.py press              # recent press mentions
python3 tools/clis/ransomwarelive.py press --all        # full archive
python3 tools/clis/ransomwarelive.py sectors            # valid sector filter values
python3 tools/clis/ransomwarelive.py csirt NL           # CSIRT contacts for a country

All commands accept --dry-run (preview the request without spending quota) and --insecure (TLS bypass — rarely needed). The CLI exits 2 if RANSOMWARE_LIVE is unset (not in dry-run); report missing key, do not fabricate.

Quota

PRO tier: 3,000 calls/day with burst allowed. Each subcommand is one HTTP call. Use --dry-run if scoping a batch.

Pivots

  • From /domain-investigation or /ip-investigation → search --q <orgname> to check if the domain's owner has been claimed
  • From /ransomware-ecosystem knowledge cell → groups --limit 30 for current top-N + group-profile for each
  • From /threat-actor-profiling (ransomware group target) → group-profile <name> is the primary feed; chain iocs, yara, negotiations for depth
  • From /sigma-writing / /yara-writing → yara <group> returns existing community rules to start from
  • Country-scoped briefing → search --country <CC>, then csirt <CC> for IR contact list

Response format

Read commands return:

source: ransomware.live
operation: search | recent | victim | groups | group-profile | iocs | yara | ...
query_time: <ISO8601>
data:
  <distilled response>

For search and recent, data.victims[] is normalised to:

id: <ransomware.live id>
title: <victim name>
group: <ransomware group, e.g. lockbit3, akira>
country: <ISO-3166 alpha-2>
sector: <activity / sector>
discovered: <when ransomware.live scraped the post>
published: <when leak-site authored the post>
description: <criminal-written breach summary — treat as low credibility>
website: <victim website>
post_url: <leak-site post URL, often .onion>
permalink: <ransomware.live record URL>
screenshot: <leak-site screenshot, if archived>

Source reliability (Admiralty default)

Default for /score-source: B2 (usually reliable, probably true).

Important credibility nuance:

FieldCredibilityWhy
Victim name, country, sector, dates, group attribution2 (probably true)Corroborated by the leak-site post itself
Breach description3–4 (possibly true → doubtful)Written by criminals to coerce payment; routinely overstates volume/sensitivity/ongoing access
YARA rules, IOC lists2 (probably true)Sourced from public researchers; cross-check before deploying

A leak-site claim is a claim, not a confirmed breach. Many victims dispute or never publicly acknowledge. Some posts repackage older breaches under a new brand. Always frame downstream products with this caveat.

Operational notes

  • Server returns full result sets, not pages. search --country US hits the API once and returns ~8k victims. --limit N trims locally — narrow with filters (--group, --sector, --country) to keep responses small.
  • group_name is sometimes null in /victims/recent responses — known API quirk; the value populates correctly in search and victim detail.
  • The description field can contain victim PII or stolen credentials — quote selectively in finished products; don't paste full responses into shared docs.
  • PRO endpoint is api-pro.ransomware.live. The free unauthenticated endpoint (api.ransomware.live, 1 req/min) is not supported by this CLI.

Related skills

  • /ransomware-ecosystem — knowledge cell consuming groups + group-profile data
  • /threat-actor-profiling — for ransomware-group profiles, the group-profile endpoint is the primary feed
  • /yara-writing, /sigma-writing — start from yara <group> output
  • /lookup-virustotal, /lookup-otx — chain after pulling group IOCs to verify hashes
  • /domain-investigation — pivots into search --q <orgname> for victim-status check
  • /initial-access-brokers — complementary; ransomware.live doesn't track IAB activity
  • /score-source, /apply-tlp, /confidence-language — apply rigor before publishing

See also

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

AbuseIPDB API reference. IP reputation and abuse report lookups.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Liberty91LTD/cti-skills272026년 9월 30일 업데이트

ach

무료

Analysis of Competing Hypotheses — structured technique for evaluating multiple explanations against evidence. Use when facing ambiguous attribution or multiple plausible scenarios. Prefers graded evidence items from /quality-of-information-check and runs that skill first when handed raw URLs or text. Runs on ungraded evidence when asked, labelled as such, unweighted, and capped at Moderate confidence.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Liberty91LTD/cti-skills272026년 9월 30일 업데이트

Use when documenting a named campaign across time and victims, the user asks to start or update a campaign record, or another skill identified a multi-incident cluster that warrants formal tracking. Provides the template (timeline, attribution, victimology, attack chain, Diamond Model mapping, IOC clusters) and the lifecycle from active to historical.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Liberty91LTD/cti-skills272026년 9월 30일 업데이트

Use when the user asks about carding, BIN attacks, payment-card breach markets, fullz/CVV2 trade, autoshops (BidenCash, Brian's Club, Russianmarket, B1ack's Stash), or financial-fraud TTPs. Self-updating knowledge cell.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Liberty91LTD/cti-skills272026년 9월 30일 업데이트

Censys API v2 reference. Host reconnaissance and certificate data.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Liberty91LTD/cti-skills272026년 9월 30일 업데이트

Use when the user asks about Chinese state-sponsored cyber operations or specific PRC-aligned actors (APT41, Volt Typhoon, Mustang Panda, APT10, APT31, Salt Typhoon, etc.), MSS/PLA-attributed campaigns, or PRC sector targeting. Self-updating knowledge cell.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Liberty91LTD/cti-skills272026년 9월 30일 업데이트

Liberty91LTD의 스킬 모두 보기

이 스킬의 문제 신고