Héctor Jara

#AI #Data #Security #crypto #governance #ISO27001 #PCI #privacy #cloud #educacion #Sustainability | fan y socio @RacingClub | work @VerSprite

[VerSprite Cybersecurity Services @VerSprite]

Last day at EkoParty! Our CTF wraps up at 16:00. Come finish what you started or just swing by to say hello before everyone packs up. #ekoparty

[VerSprite Cybersecurity Services @VerSprite]

We're heading to Buenos Aires for EkoParty 🎉
Oct 7-9, the VerSprite team is there in person, running a CTF. Come test your skills against us and the community🚩
If you're at EkoParty, come say hi!

[hernannadal @hernannadal]

A mi la IA ….

[VerSprite Cybersecurity Services @VerSprite]

Your IT team didn't call asking you to "update your passkey."

Microsoft has flagged an active campaign impersonating IT support to hijack Microsoft 365 accounts via fake passkey/MFA/SSO updates.

Never approve an MFA prompt you didn't request. Verify through your normal IT channel.

[VerSprite Cybersecurity Services @VerSprite]

Threat Modeling as a Service (TMaaS) delivers measurable operational and security improvements:

Scalability to match fluctuating demand without permanent headcount

Consistency through standardized threat assessment

Specialized external perspective that reduces institutional bias

Access to current threat intelligence for comprehensive coverage

By outsourcing threat modeling, development teams stay focused on core functionality while security champions ensure gaps are closed before deployment.

Learn more: https://t.co/VGz2DzPEkA

#CyberThreatModeling #InfoSec #DevSecOps #ZeroTrust

[VerSprite Cybersecurity Services @VerSprite]

Vulnerability Disclosure: Prompt Injection in CodeGPT

Ramiro uncovered a prompt injection vulnerability affecting Visual Studio Code extension, CodeGPT in the way it processes markdown, allowing attackers to exfiltrate sensitive data through external image requests.

LLMs embedded in developer environments must be treated as semi-trusted agents. Without strict input validation and output controls, they become viable attack surfaces.

Read the full disclosure: https://t.co/0ihEMSIrLs

#VulnerabilityDisclosure #PromptInjection #LLMSecurity #SecureAI #CyberSecurity

[VerSprite Cybersecurity Services @VerSprite]

LLMs and Document Poisoning

Legal firms are adopting LLMs for contract review, case summarization, and discovery automation.
But prompt injection introduces a new class of risk: document poisoning.

VerSprite’s research shows how adversarial clauses embedded in legal documents can manipulate LLM outputs, misinterpret intent, or even suppress critical findings.

This undermines the reliability of AI-assisted legal workflows and introduces liability in client representation.

If your firm uses LLMs for document analysis, you must treat prompt injection as a threat to legal accuracy and ethical compliance.

Read the full report:
https://t.co/TjkEZ9AX7q

Discover how prompt injection affects legal AI systems

#LegalTech #PromptInjection #LLMsecurity #ContractAI #DocumentPoisoning #CyberRisk #AIethics #VerSprite #LegalAI #AIvulnerabilities

[Lucas Corsino @LuccasCorsino]

🟣 Alpine F1 está buscando un Científico de Datos (Data Scientist) para trabajar en modelos de Machine Learning

Buscan alguien que logre entender la compleja data de un F1 y generar/mantener soluciones basadas en datos

Ganas de hacer patria no me faltan, pero trabajar en un equipo de F1 no es de mi agrado, ni mi objetivo personal

Si alguien más cumple los requisitos, la búsqueda está vigente en el portal de empleos de Alpine F1 (van a necesitar la visa de Global Talent de UK - leader in digital technology)

[VerSprite Cybersecurity Services @VerSprite]

AI automation platforms are rapidly becoming the backbone of enterprise workflows. But with great flexibility comes great exposure.

Most security programs are not equipped to handle the dynamic nature of LLM-driven orchestration tools.

The real risk lies in the implicit trust granted to plugins, nodes, and external APIs. Without strict validation and sandboxing, these components become vectors for data leakage, unauthorized execution, and model manipulation.

Security teams must evolve their threat models to account for AI-native risks. Traditional perimeter defenses won’t stop a poisoned vector database or a prompt injection that rewires business logic.

Read the full analysis: https://t.co/3Xeh6Fn7ce

Learn how to secure your AI workflows before they become your adversary’s playground.

#AIWorkflowSecurity #LLMThreats #CyberRisk #AutomationSecurity #PromptInjection #DataExfiltration #SecurityEngineering

[Racing Club @RacingClub]

Mire el estadio de La Academia, el Coliseo siempre está de fiesta 🎶

¡Felices 7️⃣5️⃣ Cilindro, por muchas velitas más!

🎙️ Gracias @casciari por tu voz.

[VerSprite Cybersecurity Services @VerSprite]

Structured prompting isn’t a silver bullet.

Once content flows through translation or markdown, wrappers degrade and raw instructions resurface.

Full breakdown:
https://t.co/VoYDIKxA8P

#AIExploitation #PromptInjection #LLMRedTeam #Cybersecurity

[VerSprite Cybersecurity Services @VerSprite]

Privilege escalation isn't always about the shell.
Kernel memory leaks can offer direct paths to root access.

Details:
https://t.co/3RzU6YsRwN

#PrivilegeEscalation #LinuxSecurity #ExploitDev

[VerSprite Cybersecurity Services @VerSprite]

PASTA is not just a framework. It’s how we align business risk with technical threats.
VerSprite’s TMaaS operationalizes PASTA across your SDLC.

See how: https://t.co/NDGbkTUIBg

#PASTA #ThreatModeling #CyberRisk

[VerSprite Cybersecurity Services @VerSprite]

CREST/CIS Accreditation isn’t given—it’s earned.
VerSprite’s DevSecOps and cloud security practices are now globally recognized.

Discover our secure-by-design approach:
🔗 https://t.co/vR2XZ1Ntqm

#DevSecOps #CRESTAccredited #CloudSecurity #CyberResilience #SecurityEngineering

[VerSprite Cybersecurity Services @VerSprite]

FORK maps CVEs to EPSS and real-world TTPs.
Threat modeling becomes actionable and intelligence-driven.

https://t.co/XSB8q7o1BG
#AppSec #ThreatIntel #OWASP #CyberDefense

[VerSprite Cybersecurity Services @VerSprite]

From cloud to legacy, VerSprite applies CIS Controls to reduce risk with precision. CREST accreditation ensures our assessments are globally benchmarked and technically sound.
🔗https://t.co/vR2XZ1Ntqm
#CyberSecurity #CREST #CISControls #RiskManagement

[VerSprite Cybersecurity Services @VerSprite]

Stage 6 of PASTA: Simulate real attacks. JWT forgery, NoSQL injection, privilege escalation. If you’re not testing like an attacker, you’re not testing.
#APIsecurity #RedTeamOps #PenTest #CyberDefense https://t.co/zJEmL6t4Xd

[VerSprite Cybersecurity Services @VerSprite]

Shadow AI = governance gap. AI tools are being adopted without IT, risk, or compliance oversight. Centralize AI governance now. #ShadowAI #AIGovernance #CISO #CyberSecurity

🔗 https://t.co/fvU5irhFUy

[VerSprite Cybersecurity Services @VerSprite]

Shadow AI = unmanaged AI tools in your org. It’s not innovation—it’s exposure. Learn how to detect and contain it before it becomes a breach.

#CyberSecurity #ShadowAI #AIrisks

Read more: https://t.co/fvU5irhFUy

[VerSprite Cybersecurity Services @VerSprite]

Southeast Asia is emerging as a cybercrime innovation hub.

AI + traditional crime = a new hybrid threat model.
📄 https://t.co/L2kRTiy2WY

#CyberCrime #ThreatLandscape #Infosec #CTI #GlobalSecurity