Datasets:

emgena
/

captcha_adversarial_vlm_loop_terminator_teaser

Official VLM & GUI Computer-Use Agent Guard Teaser by Emgena AI Labs.
Contains 50 verified production trajectories for DevSecOps & Cloud Security.
Hol dir das vollständige Turnkey-Paket (500 Samples + AST MCP Guard) direkt bei EMGENA AI Labs
(Gutscheincode LAUNCH20 für 20 € Direktrabatt & sofortigen Download)

Enterprise USP & Features

Turnkey Enterprise Model Context Protocol (MCP) Guard for Cursor IDE & Claude Desktop. Protects against Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety.

Deterministic In-IDE Guardrails: Validated across Cursor IDE and Claude Desktop.
100% AST Syntactic Validation (Zero Syntax Errors)
EU AI Act & EULA Compliance: Certified for enterprise deployment under Articles 50 & 53.

Full Enterprise MCP Suite

Die vollständige Version enthält 500 AST-geprüfte Trajektorien, 30 Produktions-Incidents, den Zero-Dependency MCP Server (server.py), DPO-Präferenzpaare und BGB/EU AI Act Compliance.

Vollständiges Paket kaufen: EMGENA AI Labs Store
20 € Rabatt-Gutschein: LAUNCH20
Offizieller Webshop: https://emgena.com/trainingslager

Downloads last month: 45

Number of rows:

Total file size:

1.13 MB

id stringlengths 49 49	incident_id stringlengths 47 47	domain stringclasses 1 value	messages listlengths 5 5
captcha_adversarial_vlm_loop_terminator_traj_0000	captcha_adversarial_vlm_loop_terminator_inc_001	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are Emgena AI Labs Enterprise Sentinel for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Analyze production telemetry, trace defect root cause, invoke diagnostic tools, and deliver AST-validated zero-defect remediations." }, { "role": "use...
captcha_adversarial_vlm_loop_terminator_traj_0001	captcha_adversarial_vlm_loop_terminator_inc_002	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are the FAANG v2.5 Staff Reliability Architect specializing in Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Inspect runtime traces and execute strict AST static analysis guards to protect critical production infrastructure." }, { "role": ...
captcha_adversarial_vlm_loop_terminator_traj_0002	captcha_adversarial_vlm_loop_terminator_inc_003	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Autonomous AI Defense Guard for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Enforce zero-tolerance safety policies, eliminate memory leaks and runtime deadlocks, and verify code invariance via AST parsing." }, { "role": "user", ...
captcha_adversarial_vlm_loop_terminator_traj_0003	captcha_adversarial_vlm_loop_terminator_inc_004	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Enterprise SRE Sentinel operating under strict EU AI Act compliance. Diagnose incidents, execute diagnostic tools, and output production-ready remediations." }, { "role": "user", "content": "CRITICAL PRODUCTION ALERT in service 'captcha_adversarial_vlm_...
captcha_adversarial_vlm_loop_terminator_traj_0004	captcha_adversarial_vlm_loop_terminator_inc_005	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are Emgena AI Labs Enterprise Sentinel for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Analyze production telemetry, trace defect root cause, invoke diagnostic tools, and deliver AST-validated zero-defect remediations." }, { "role": "use...
captcha_adversarial_vlm_loop_terminator_traj_0005	captcha_adversarial_vlm_loop_terminator_inc_006	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are the FAANG v2.5 Staff Reliability Architect specializing in Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Inspect runtime traces and execute strict AST static analysis guards to protect critical production infrastructure." }, { "role": ...
captcha_adversarial_vlm_loop_terminator_traj_0006	captcha_adversarial_vlm_loop_terminator_inc_007	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Autonomous AI Defense Guard for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Enforce zero-tolerance safety policies, eliminate memory leaks and runtime deadlocks, and verify code invariance via AST parsing." }, { "role": "user", ...
captcha_adversarial_vlm_loop_terminator_traj_0007	captcha_adversarial_vlm_loop_terminator_inc_008	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Enterprise SRE Sentinel operating under strict EU AI Act compliance. Diagnose incidents, execute diagnostic tools, and output production-ready remediations." }, { "role": "user", "content": "CRITICAL PRODUCTION ALERT in service 'captcha_adversarial_vlm_...
captcha_adversarial_vlm_loop_terminator_traj_0008	captcha_adversarial_vlm_loop_terminator_inc_009	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are Emgena AI Labs Enterprise Sentinel for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Analyze production telemetry, trace defect root cause, invoke diagnostic tools, and deliver AST-validated zero-defect remediations." }, { "role": "use...
captcha_adversarial_vlm_loop_terminator_traj_0009	captcha_adversarial_vlm_loop_terminator_inc_010	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are the FAANG v2.5 Staff Reliability Architect specializing in Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Inspect runtime traces and execute strict AST static analysis guards to protect critical production infrastructure." }, { "role": ...
captcha_adversarial_vlm_loop_terminator_traj_0010	captcha_adversarial_vlm_loop_terminator_inc_011	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Autonomous AI Defense Guard for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Enforce zero-tolerance safety policies, eliminate memory leaks and runtime deadlocks, and verify code invariance via AST parsing." }, { "role": "user", ...
captcha_adversarial_vlm_loop_terminator_traj_0011	captcha_adversarial_vlm_loop_terminator_inc_012	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Enterprise SRE Sentinel operating under strict EU AI Act compliance. Diagnose incidents, execute diagnostic tools, and output production-ready remediations." }, { "role": "user", "content": "CRITICAL PRODUCTION ALERT in service 'captcha_adversarial_vlm_...
captcha_adversarial_vlm_loop_terminator_traj_0012	captcha_adversarial_vlm_loop_terminator_inc_013	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are Emgena AI Labs Enterprise Sentinel for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Analyze production telemetry, trace defect root cause, invoke diagnostic tools, and deliver AST-validated zero-defect remediations." }, { "role": "use...
captcha_adversarial_vlm_loop_terminator_traj_0013	captcha_adversarial_vlm_loop_terminator_inc_014	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are the FAANG v2.5 Staff Reliability Architect specializing in Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Inspect runtime traces and execute strict AST static analysis guards to protect critical production infrastructure." }, { "role": ...
captcha_adversarial_vlm_loop_terminator_traj_0014	captcha_adversarial_vlm_loop_terminator_inc_015	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Autonomous AI Defense Guard for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Enforce zero-tolerance safety policies, eliminate memory leaks and runtime deadlocks, and verify code invariance via AST parsing." }, { "role": "user", ...
captcha_adversarial_vlm_loop_terminator_traj_0015	captcha_adversarial_vlm_loop_terminator_inc_016	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Enterprise SRE Sentinel operating under strict EU AI Act compliance. Diagnose incidents, execute diagnostic tools, and output production-ready remediations." }, { "role": "user", "content": "CRITICAL PRODUCTION ALERT in service 'captcha_adversarial_vlm_...
captcha_adversarial_vlm_loop_terminator_traj_0016	captcha_adversarial_vlm_loop_terminator_inc_017	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are Emgena AI Labs Enterprise Sentinel for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Analyze production telemetry, trace defect root cause, invoke diagnostic tools, and deliver AST-validated zero-defect remediations." }, { "role": "use...
captcha_adversarial_vlm_loop_terminator_traj_0017	captcha_adversarial_vlm_loop_terminator_inc_018	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are the FAANG v2.5 Staff Reliability Architect specializing in Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Inspect runtime traces and execute strict AST static analysis guards to protect critical production infrastructure." }, { "role": ...
captcha_adversarial_vlm_loop_terminator_traj_0018	captcha_adversarial_vlm_loop_terminator_inc_019	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Autonomous AI Defense Guard for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Enforce zero-tolerance safety policies, eliminate memory leaks and runtime deadlocks, and verify code invariance via AST parsing." }, { "role": "user", ...
captcha_adversarial_vlm_loop_terminator_traj_0019	captcha_adversarial_vlm_loop_terminator_inc_020	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Enterprise SRE Sentinel operating under strict EU AI Act compliance. Diagnose incidents, execute diagnostic tools, and output production-ready remediations." }, { "role": "user", "content": "CRITICAL PRODUCTION ALERT in service 'captcha_adversarial_vlm_...
captcha_adversarial_vlm_loop_terminator_traj_0020	captcha_adversarial_vlm_loop_terminator_inc_021	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are Emgena AI Labs Enterprise Sentinel for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Analyze production telemetry, trace defect root cause, invoke diagnostic tools, and deliver AST-validated zero-defect remediations." }, { "role": "use...
captcha_adversarial_vlm_loop_terminator_traj_0021	captcha_adversarial_vlm_loop_terminator_inc_022	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are the FAANG v2.5 Staff Reliability Architect specializing in Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Inspect runtime traces and execute strict AST static analysis guards to protect critical production infrastructure." }, { "role": ...
captcha_adversarial_vlm_loop_terminator_traj_0022	captcha_adversarial_vlm_loop_terminator_inc_023	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Autonomous AI Defense Guard for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Enforce zero-tolerance safety policies, eliminate memory leaks and runtime deadlocks, and verify code invariance via AST parsing." }, { "role": "user", ...
captcha_adversarial_vlm_loop_terminator_traj_0023	captcha_adversarial_vlm_loop_terminator_inc_024	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Enterprise SRE Sentinel operating under strict EU AI Act compliance. Diagnose incidents, execute diagnostic tools, and output production-ready remediations." }, { "role": "user", "content": "CRITICAL PRODUCTION ALERT in service 'captcha_adversarial_vlm_...
captcha_adversarial_vlm_loop_terminator_traj_0024	captcha_adversarial_vlm_loop_terminator_inc_025	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are Emgena AI Labs Enterprise Sentinel for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Analyze production telemetry, trace defect root cause, invoke diagnostic tools, and deliver AST-validated zero-defect remediations." }, { "role": "use...
captcha_adversarial_vlm_loop_terminator_traj_0025	captcha_adversarial_vlm_loop_terminator_inc_026	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are the FAANG v2.5 Staff Reliability Architect specializing in Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Inspect runtime traces and execute strict AST static analysis guards to protect critical production infrastructure." }, { "role": ...
captcha_adversarial_vlm_loop_terminator_traj_0026	captcha_adversarial_vlm_loop_terminator_inc_027	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Autonomous AI Defense Guard for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Enforce zero-tolerance safety policies, eliminate memory leaks and runtime deadlocks, and verify code invariance via AST parsing." }, { "role": "user", ...
captcha_adversarial_vlm_loop_terminator_traj_0027	captcha_adversarial_vlm_loop_terminator_inc_028	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Enterprise SRE Sentinel operating under strict EU AI Act compliance. Diagnose incidents, execute diagnostic tools, and output production-ready remediations." }, { "role": "user", "content": "CRITICAL PRODUCTION ALERT in service 'captcha_adversarial_vlm_...
captcha_adversarial_vlm_loop_terminator_traj_0028	captcha_adversarial_vlm_loop_terminator_inc_029	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are Emgena AI Labs Enterprise Sentinel for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Analyze production telemetry, trace defect root cause, invoke diagnostic tools, and deliver AST-validated zero-defect remediations." }, { "role": "use...
captcha_adversarial_vlm_loop_terminator_traj_0029	captcha_adversarial_vlm_loop_terminator_inc_030	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are the FAANG v2.5 Staff Reliability Architect specializing in Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Inspect runtime traces and execute strict AST static analysis guards to protect critical production infrastructure." }, { "role": ...
captcha_adversarial_vlm_loop_terminator_traj_0030	captcha_adversarial_vlm_loop_terminator_inc_001	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Autonomous AI Defense Guard for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Enforce zero-tolerance safety policies, eliminate memory leaks and runtime deadlocks, and verify code invariance via AST parsing." }, { "role": "user", ...
captcha_adversarial_vlm_loop_terminator_traj_0031	captcha_adversarial_vlm_loop_terminator_inc_002	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Enterprise SRE Sentinel operating under strict EU AI Act compliance. Diagnose incidents, execute diagnostic tools, and output production-ready remediations." }, { "role": "user", "content": "CRITICAL PRODUCTION ALERT in service 'captcha_adversarial_vlm_...
captcha_adversarial_vlm_loop_terminator_traj_0032	captcha_adversarial_vlm_loop_terminator_inc_003	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are Emgena AI Labs Enterprise Sentinel for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Analyze production telemetry, trace defect root cause, invoke diagnostic tools, and deliver AST-validated zero-defect remediations." }, { "role": "use...
captcha_adversarial_vlm_loop_terminator_traj_0033	captcha_adversarial_vlm_loop_terminator_inc_004	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are the FAANG v2.5 Staff Reliability Architect specializing in Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Inspect runtime traces and execute strict AST static analysis guards to protect critical production infrastructure." }, { "role": ...
captcha_adversarial_vlm_loop_terminator_traj_0034	captcha_adversarial_vlm_loop_terminator_inc_005	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Autonomous AI Defense Guard for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Enforce zero-tolerance safety policies, eliminate memory leaks and runtime deadlocks, and verify code invariance via AST parsing." }, { "role": "user", ...
captcha_adversarial_vlm_loop_terminator_traj_0035	captcha_adversarial_vlm_loop_terminator_inc_006	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Enterprise SRE Sentinel operating under strict EU AI Act compliance. Diagnose incidents, execute diagnostic tools, and output production-ready remediations." }, { "role": "user", "content": "CRITICAL PRODUCTION ALERT in service 'captcha_adversarial_vlm_...
captcha_adversarial_vlm_loop_terminator_traj_0036	captcha_adversarial_vlm_loop_terminator_inc_007	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are Emgena AI Labs Enterprise Sentinel for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Analyze production telemetry, trace defect root cause, invoke diagnostic tools, and deliver AST-validated zero-defect remediations." }, { "role": "use...
captcha_adversarial_vlm_loop_terminator_traj_0037	captcha_adversarial_vlm_loop_terminator_inc_008	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are the FAANG v2.5 Staff Reliability Architect specializing in Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Inspect runtime traces and execute strict AST static analysis guards to protect critical production infrastructure." }, { "role": ...
captcha_adversarial_vlm_loop_terminator_traj_0038	captcha_adversarial_vlm_loop_terminator_inc_009	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Autonomous AI Defense Guard for Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety. Enforce zero-tolerance safety policies, eliminate memory leaks and runtime deadlocks, and verify code invariance via AST parsing." }, { "role": "user", ...
captcha_adversarial_vlm_loop_terminator_traj_0039	captcha_adversarial_vlm_loop_terminator_inc_010	Vision-Language Models (VLM), GUI Computer-Use & Desktop Agent Safety	[ { "role": "system", "content": "You are an Enterprise SRE Sentinel operating under strict EU AI Act compliance. Diagnose incidents, execute diagnostic tools, and output production-ready remediations." }, { "role": "user", "content": "CRITICAL PRODUCTION ALERT in service 'captcha_adversarial_vlm_...