# PROMPTFLUX Gemini API self-modifying VBScript dropper

- ID: promptflux-gtig-2025-06
- Reported: 2025-11-05
- Researcher: Google Threat Intelligence Group
- Organization: Google GTIG
- AI role: llm_generated_malware (LLM-generated malware)
- Confidence in AI involvement: not assessed
- Research / borderline: true
- Last reviewed: 2026-10-20
- Entry page: https://openattacks.ai/attacks/promptflux-gtig-2025-06/
- Entry JSON: https://openattacks.ai/api/attacks/promptflux-gtig-2025-06.json
- Browse title: PROMPTFLUX Gemini-powered self-modifying VBScript dropper
- Editorial summary: An experimental VBScript dropper queries Gemini for antivirus-evasion rewrites. The analyzed self-update function was commented out, so active self-rewriting is not established.
- Evidence class (catalog annotation): Malware analysis
- Source-stated target environments: Not specified
- Stage hunts and conditional defenses: https://openattacks.ai/api/defenses/promptflux-gtig-2025-06.json

## Evidence qualification
GTIG classifies as experimental/dev; incomplete features; no demonstrated ability to compromise victims.

## Technique
VBScript dropper with decoy installer; 'Thinking Robot' module POSTs to Gemini API (gemini-1.5-flash-latest) with hard-coded key asking for VBScript AV-evasion rewrites; intends Startup persistence and removable-drive spread. Self-update function commented out in analyzed samples.

## Stack relevance (editorial)
Assessment: attacker\_tool\_only. Editorial applicability reviewed against the catalog source account; not a scan or independent incident verification.
The named AI product is used by the attacker. No reviewed mapping from owning that product to the reported victim conditions is available.
A product mention alone is not a victim-side applicability mapping. Selection does not verify affected versions, permissions or compromise.

## Initial access
unknown

## AI tooling and models
- [Google Gemini API](https://ai.google.dev/)
- [gemini-1.5-flash-latest](https://ai.google.dev/)

## Other tooling
- VBScript
- Startup folder persistence

## Victims and sectors
unknown

## Target geography
- unknown

## Attribution
unattributed; financially motivated lure filenames suggested

## Scale and metrics
Experimental; Google disabled associated assets

## Attack flow
Catalog reconstruction of the supplied source account. Unknowns and research caveats are preserved.
### step-1: Initial delivery is not established [unknown]
The analyzed VBScript includes a decoy installer, but the initial route is unknown.
Evidence: https://cloud.google.com/blog/topics/threat-intelligence/threat-actor-usage-of-ai-tools/

### step-2: Thinking Robot queries Gemini [ai]
A hard-coded key supports requests for VBScript evasion rewrites.
Evidence: https://cloud.google.com/blog/topics/threat-intelligence/threat-actor-usage-of-ai-tools/

### step-3: Dropper contains a rewrite workflow [execution]
The analyzed sample contains code for self-modification.
Evidence: https://cloud.google.com/blog/topics/threat-intelligence/threat-actor-usage-of-ai-tools/

### step-4: Self-update is commented out [unknown]
Startup persistence and removable-drive spread are intended behavior; the analyzed rewrite function was not active.
Evidence: https://cloud.google.com/blog/topics/threat-intelligence/threat-actor-usage-of-ai-tools/

Connections:
- step-1 -> step-2
- step-2 -> step-3
- step-3 -> step-4

## Stage defense and hunting recommendations
Editorial recommendations, not validated coverage. Source-controlled query starters are drafts. Custom-schema examples require telemetry collection and field mapping; absence of a mapping means not assessed. MagicSword policy opportunities require managed Windows scope and enforcement.
### step-1: No source-backed runtime hunt yet
- Visibility: Not established
- Query origin: catalog_authored; status: not_applicable; log scope: gap
- Hunt hypothesis: The supplied account does not establish a deployed victim-side behavior for this stage.
- Required telemetry: not assessed
- Investigate: Obtain a source-backed procedure and identify which party could observe it before drafting a detection.
- Limitation: Marketing claims, unknown steps and unconfirmed deployments are not detection coverage. No rule is generated for this stage.
- Qualification: The supplied account leaves this stage unestablished. Obtain a source-backed procedure before proposing a runtime query.
- No direct MagicSword prevention claim is recorded for this stage.

### step-2: Unexpected process → model service
- Visibility: Defender-side logs where collected
- Query origin: catalog_authored; status: draft; log scope: custom_schema
- SIGMA query download: /api/detections/model-network.sigma.yml
- KQL query download: /api/detections/model-network.kql
- SPL query download: /api/detections/model-network.spl
- Hunt hypothesis: An unexpected script, implant or local model client communicates with a model service near command execution, collection or file modification.
- Required telemetry: Per-process network / DNS records; AI-service request logs where available; Process and file behavior on the calling host
- Investigate: Join the network connection to the process tree and the actions before and after the response.
- Investigate: Compare with approved AI integrations; inspect local model listeners for PoCs that do not use a hosted API.
- Limitation: Connecting to a legitimate model domain does not establish malicious use or reveal returned code. Local / in-process models can evade network-only hunts.
- No direct MagicSword prevention claim is recorded for this stage.

### step-3: Suspicious Windows script execution
- Visibility: Defender-side logs where collected
- Query origin: catalog_authored; status: draft; log scope: native
- SIGMA query download: /api/detections/windows-shell.sigma.yml
- KQL query download: /api/detections/windows-shell.kql
- SPL query download: /api/detections/windows-shell.spl
- Hunt hypothesis: A script host or shell runs encoded content, a remote URL, or inline script syntax, then downloads or launches a payload.
- Required telemetry: Windows process creation: Sysmon Event 1 or Defender DeviceProcessEvents; PowerShell script-block logging when enabled; Download / network and installer events
- Investigate: Follow the parent and child process tree and recover the script or URL from logs without executing it.
- Investigate: Correlate the payload's signer/hash, persistence, and network activity; baseline approved automation.
- Limitation: Encoded commands and remote administration can be legitimate. This does not cover every script host, in-process execution, or non-Windows branch.
- Conditional MagicSword policy: Use MagicSword to review and enforce a policy restricting unneeded Windows script hosts, shells or installer utilities identified in this branch.
- Policy scope: Windows VBScript execution. The analyzed sample's self-update function was commented out; blocking it is not proof of stopping an active self-rewrite chain.
- Policy limitation: Policy opportunity, not tested incident coverage. Requires enrolled Windows devices, a policy covering the actual executable, and enforcement. Audit mode records decisions without blocking; allowed tools and in-process / cloud actions can remain reachable.
- Policy reference: https://www.magicsword.io/use-cases/application-control?utm_source=openattacks&utm_medium=referral&utm_content=stage-control

### step-4: No source-backed runtime hunt yet
- Visibility: Not established
- Query origin: catalog_authored; status: not_applicable; log scope: gap
- Hunt hypothesis: The supplied account does not establish a deployed victim-side behavior for this stage.
- Required telemetry: not assessed
- Investigate: Obtain a source-backed procedure and identify which party could observe it before drafting a detection.
- Limitation: Marketing claims, unknown steps and unconfirmed deployments are not detection coverage. No rule is generated for this stage.
- Qualification: The supplied account leaves this stage unestablished. Obtain a source-backed procedure before proposing a runtime query.
- No direct MagicSword prevention claim is recorded for this stage.


## MITRE ATLAS
Mappings are catalog annotations based on described behavior, not assignments or endorsements by MITRE.
- AML.T0102: Generate Malicious Commands — A dropper requests generated VBScript rewrites; the analyzed self-update routine is commented out.
  Technique: https://atlas.mitre.org/techniques/AML.T0102
  Evidence: https://cloud.google.com/blog/topics/threat-intelligence/threat-actor-usage-of-ai-tools/

## Defensive notes
Monitor Gemini API keys embedded in scripts; block unexpected generative API from endpoints

## Sources
- https://cloud.google.com/blog/topics/threat-intelligence/threat-actor-usage-of-ai-tools/
