Suspicious
New PDF undetected by engines but shows suspicious sandbox behaviors like process injection, LSASS access, and direct IP connections.
d905f6f48cbb8bb8d8…1179dab558The verdict, reasoned out.
Not a rules engine. The MT AI Engine reads every signal we collected, weighs them against history, and commits to an answer.
Static analysis is entirely clean with zero malicious or suspicious flags from good engine coverage. Dynamic sandbox behavior triggers high-confidence synthesis rules for process injection into Explorer, potential LSASS credential access, and direct-IP C2 patterns typical of exploits. The PDF nature and business-proposal filename align with phishing vectors, but lack of sandbox malicious verdict, unknown dropped child, and unflagged IPs temper the concern. No historical RAG, external intel, or feedback provides additional context. Overall, dynamic risks outweigh static cleanliness for a suspicious call.
Each signal cites a concrete token from the evidence the arbiter saw — engine name, MITRE technique, signer string, or an exact count.
0/63 engines malicious (17 tier1 clean)
triggeredHeuristics 'MalwareTips.Synth.ProcessInjection' (high, T1055)
behaviour.contactedIps length=3 (184.29.30.201 etc.), no contactedDomains
file.ageDays=1, prevalence.classification='rare_new'
behaviour.offensiveCount=3 including T1003/T1055
- 0 malicious from 63 engines (17 tier1 clean)
- No malicious sandbox verdict
- No malicious dropped children
- No external intel hits
- No hacktool/PUA labels
- Process injection heuristic (T1055, high severity)
- Credential dumping pattern (T1003, LSASS access)
- Direct IP connections without DNS (3 IPs)
- Rare new prevalence (1 submission, 1 day old)
- PDF file type (common exploit vector)
- Offensive MITRE techniques (count=3)
Quarantine this file and avoid opening it. Perform a full system scan with updated security tools and analyze the dropped child file if present.
What this file did when executed
This file was detonated in 1 sandbox and its runtime behaviour was observed.
Adversary techniques mapped to the MITRE ATT&CK framework.
- 184.29.30.201
- 23.22.254.206
- 162.159.36.2
- C:\Users\<USER>\AppData\Local\Temp\acroNGLLog.txt
- C:\Users\<USER>\AppData\Local\Temp\NGL\
- C:\Users\<USER>\AppData\Local\Temp\TmpE92.tmp
- C:\Users\<USER>\AppData\Local\Temp\A9qjy75q_8vkj4l_4x8.tmp
- C:\Windows\ServiceProfiles\LocalService\AppData\Local\FontCache\Fonts\Download-1.tmp
- C:\Users\<USER>\AppData\Roaming\Adobe\Acrobat\DC\JSCache\GlobSettings
- C:\Users\<USER>\AppData\Local\Packages\Microsoft.XboxGamingOverlay_8wekyb3d8bbwe\LocalCache\KnownGameList.bin
- C:\Users\<USER>\AppData\Local\Microsoft\GameDVR\KnownGameList.update
- Local\SyncServiceThread
- Local\SessionImmersiveColorMutex
- Global\_MSIExecute
- Global\MSILOG_4ce7be391dce1b5GOL.b1211ISM_pmeT_lacoL_ataDppA_onurB_sresU_:C
- Global\AdobeCrashProcessorLocalLowLock
Files this sample writes at runtime
This file drops 1 child at runtime. None are currently flagged malicious in our cache.
- a779a261df447a4c298c…b1b86dNever scannednever seen before
YARA + heuristic rules that fired
A researcher-curated or high-severity heuristic rule matched this sample. These rules target specific malware families and are near-definitive.
MITRE T1055 (Process Injection) observed — CreateRemoteThread / APC / reflective-DLL injection. The payload is being smuggled into a legitimate process to bypass AV hooks.
EvidenceC:\Windows\Explorer.EXESandbox observed process activity targeting LSASS (Windows credential store). Legitimate software has no business reading LSASS memory — this is Mimikatz-shape behaviour.
EvidenceC:\Windows\system32\lsass.exeSample contacted 3 external IP address(es) and zero domains. Benign software virtually always uses DNS; no-DNS direct-IP C2 is a strong malware indicator because it bypasses reputation systems and dodges domain-based blocklists.
Evidence184.29.30.201 · 23.22.254.206 · 162.159.36.2
0 detections across 75 engines
How often this file shows up in the wild
Barely seen in the wild and first surfaced recently. This is the footprint of targeted malware the AV industry hasn't signatured yet — extra scrutiny is warranted.
Forensic fingerprint
- File name
- AlphaGraphics_Proposal_Invitation.pdf
- Size
- 102.4 KB
- MIME type
- (unknown)
- Detected type
- SHA-256
- d905f6f48cbb8bb8d8da8b2a14b09f31e23acd8c6c9fbde99a95491179dab558
- MD5
- 66257302ddfd482eeb93da32ceacacf9
- SHA-1
- eefac54390ef710172af234cca3a07b6b531fb4c
- First seen (VT)
- 5/11/2026, 2:15:11 PM
- Last analysis (VT)
- 5/11/2026, 2:15:11 PM
- First scan (MalwareTips)
- 5/12/2026, 4:57:49 PM
- Last scan (MalwareTips)
- 5/12/2026, 4:57:49 PM
Reviews & malware reports(0)
Tell the community what you saw. Tag the sample — Trojan, Adware, False Positive — and share what the file did on your system. Your report helps confirm or dispute the AV verdict.