Credential Phishing: W-2 lure with inline SVG Windows logo
Detects inbound messages containing a link with W-2 display text and an inline SVG constructed from four colored rectangles approximating the Microsoft Windows logo. Threat actors use hand-crafted SVG elements rather than image attachments to bypass image-based detection and render a convincing Windows or Microsoft brand impersonation directly in the email body. The color matching uses fuzzy hex ranges to account for minor variations across campaigns.
Sublime rule (View on GitHub)
1name: "Credential Phishing: W-2 lure with inline SVG Windows logo"
2description: "Detects inbound messages containing a link with W-2 display text and an inline SVG constructed from four colored rectangles approximating the Microsoft Windows logo. Threat actors use hand-crafted SVG elements rather than image attachments to bypass image-based detection and render a convincing Windows or Microsoft brand impersonation directly in the email body. The color matching uses fuzzy hex ranges to account for minor variations across campaigns."
3type: "rule"
4severity: "high"
5source: |
6 type.inbound
7 // display text contains a reference to W2
8 and any(body.current_thread.links,
9 regex.icontains(strings.replace_confusables(.display_text),
10 '(?:W|VV)\s*-?\s*2'
11 )
12 )
13 // inline SVG that mimics the Windows logo
14 and regex.icontains(body.html.raw,
15 '<rect x="\d+" y="\d+" width="\d+" height="\d+" fill="#[c-f][0-9a-f][0-3][0-9a-f][0-3][0-9a-f]">\s*<\/rect>\s*<rect x="\d+" y="\d+" width="\d+" height="\d+" fill="#[0-3][0-9a-f][8-f][0-9a-f][2-6][0-9a-f]">\s*<\/rect>\s*<rect x="\d+" y="\d+" width="\d+" height="\d+" fill="#[0-3][0-9a-f][5-9][0-9a-f][a-f][0-9a-f]">\s*<\/rect>\s*<rect x="\d+" y="\d+" width="\d+" height="\d+" fill="#[c-f][0-9a-f][a-f][0-9a-f][0-3][0-9a-f]">\s*<\/rect>\s*<\/svg>\s*<\/td>'
16 )
17attack_types:
18 - "Credential Phishing"
19tactics_and_techniques:
20 - "Evasion"
21 - "HTML smuggling"
22 - "Social engineering"
23detection_methods:
24 - "Content analysis"
25 - "HTML analysis"
26id: "28db8459-dbc5-52c1-8863-4ede01cf104a"