Slim Spider Steals Crypto Custody Secrets From Brazilian Financial Institution
Slim Spider is a previously undocumented financially motivated threat actor linked to attacks on Brazilian financial institutions since at least March 2026, specializing in stealing cryptocurrency custody secrets and instant payment credentials The group uses custom Bash scripts to query cloud instance metadata, steal temporary cloud credentials, and exfiltrate digital asset custody secrets, leveraging tools like Foundry's `cast` and OpenSSL for cryptographic operations Slim Spider operates a su
Analysis
TL;DR
- Slim Spider is a previously undocumented financially motivated threat actor linked to attacks on Brazilian financial institutions since at least March 2026, specializing in stealing cryptocurrency custody secrets and instant payment credentials
- The group uses custom Bash scripts to query cloud instance metadata, steal temporary cloud credentials, and exfiltrate digital asset custody secrets, leveraging tools like Foundry's
castand OpenSSL for cryptographic operations - Slim Spider operates a suite of web-based panels including NEXUS // Scanner (API endpoint scanning with Ollama), Painel de Emails Entra ID (email reconnaissance), and Painel Pix (bulk unauthorized Pix transfer execution)
- The threat actor demonstrates sophisticated cloud-native operational security, deploying backdoors mimicking legitimate infrastructure binaries and impersonating Brazil's Pix payment system through an implant named "spi"
- A parallel threat group, Breeze Comet, is also targeting Brazilian financial systems since 2024, signaling a broader shift from retail fraud to direct intrusions into core payment infrastructure across Latin America
Why It Matters
This disclosure highlights an alarming evolution in cybercrime tactics where financially motivated threat actors are developing deep cloud infrastructure expertise and targeting the credentials that sit closest to high-value digital assets. The convergence of two sophisticated groups (Slim Spider and Breeze Comet) specifically targeting Brazil's Pix instant payment system underscores the growing attractiveness of emerging payment infrastructures as attack surfaces. For AI and security practitioners, this represents a case study in how threat actors are increasingly leveraging cloud-native tooling, custom scripting, and AI-assisted reconnaissance to conduct targeted financial intrusions.
Technical Details
- Cloud Credential Theft: Slim Spider developed custom Bash scripts that query cloud instance metadata services to harvest temporary credentials over socket connections, then used the
sedcommand to clone and modify secret-extracting scripts focused on digital financial asset credentials - Cryptographic Operations: Rather than using third-party libraries that could trigger detection, the group implemented cloud-native cryptographic signing directly via OpenSSL within Bash scripts and used Foundry's
casttool to derive Ethereum wallet addresses from stolen private keys - Persistence and Evasion: The actor deployed backdoors mimicking infrastructure-related binaries to blend with legitimate tooling, pivoted through Azure DevOps to run malicious pipelines on managed Kubernetes clusters, and named an implant "spi" to impersonate Brazil's Sistema de Pagamentos Instantâneos (Pix infrastructure)
- AI-Integrated Reconnaissance: NEXUS // Scanner utilizes Ollama to categorize discovered API endpoints into 16 categories (fintech, banking, payment, cryptocurrency) and rank them based on availability and authentication options, demonstrating integration of local LLMs into offensive tooling
- Go-Based Backdoor: MikeDor, a Go-based backdoor in the group's arsenal, is capable of harvesting sensitive information and monitoring user activities, reflecting a trend toward compiled, cross-platform malware frameworks
Industry Insight
- Financial institutions operating in Latin America, particularly those integrated with Pix and similar instant payment systems, should prioritize cloud credential hygiene, metadata service hardening, and continuous monitoring of DevOps pipelines for unauthorized modifications
- The use of Ollama and local LLMs in offensive tooling (NEXUS // Scanner) signals a democratization of AI-powered reconnaissance; security teams should monitor for similar AI-integrated attack tools and update detection rules accordingly
- The shift from retail-level fraud to direct intrusions into core financial switches and payment infrastructure represents a strategic escalation that may replicate across other regions; organizations should treat payment infrastructure credentials with the same severity as cryptographic key material and implement zero-trust principles for cloud environments hosting financial assets
Disclaimer: The above content is generated by AI and is for reference only.