Original: 有關集團北美部分廠區遭網路攻擊說明
Hon Hai / Foxconn (2317) / 鴻海 (2317)Contagious Interview evolves: attackers abuse VS Code Tasks to auto-run malware when a “trusted” workspace is opened
Taiwan’s TWCERT/CC reports a technical evolution in the “Contagious Interview” campaign: instead of relying on victims to manually execute a file, attackers embed a malicious VS Code workspace configuration so code runs automatically when developers open a project folder in Trusted Mode. The technique abuses VS Code’s tasks.json automation (including a run-on-folder-open behavior) and social engineering around Workspace Trust prompts. The activity primarily targets cryptocurrency software engineers and freelancers via recruiting outreach on LinkedIn and gig platforms, then directs them to download test projects from GitHub/GitLab. TWCERT/CC says the resulting payload has been identified as a newer BeaverTail variant (Type 701), with noted functional overlap with OtterCookie (sometimes referred to as “OtterCandy”), and is focused on stealing crypto-related browser extension and wallet data as well as high-value browser-stored secrets.
Okta’s support-system intrusion highlights why HAR files and session tokens must be treated as privileged secrets
Okta’s root-cause report says a threat actor accessed files in its customer support case management system from Sept. 28 to Oct. 17, 2023, affecting 134 customers (under 1%). Some accessed files were HAR files containing session tokens, enabling session hijacking; Okta says tokens were used to hijack sessions for 5 customers. The incident stemmed from a support-system service account credential that was likely exposed after being saved to an employee’s personal Google account via Chrome sign-in on an Okta-managed laptop. Okta also disclosed a logging visibility gap that delayed identifying file downloads until an IP indicator was shared by BeyondTrust.
Microsoft’s Storm-0558 postmortem highlights identity signing-key leakage paths and validation gaps that can bridge consumer and enterprise trust domains
Microsoft’s MSRC investigation into Storm-0558 concludes that operational errors likely allowed Microsoft Account (MSA) consumer signing key material to escape a secure token signing environment via a crash-dump/debug workflow, after which the actor (attributed by Microsoft as China-based) likely obtained it by compromising a Microsoft engineer’s corporate account with access to the debugging environment. A separate engineering failure—missing issuer/scope validation when mail systems relied on a common key metadata endpoint—meant a consumer key could be used to forge tokens accepted for enterprise email access. Microsoft says it has since corrected the race condition, improved key-material detection and credential scanning, and updated libraries to automate scope validation.
CISA/FBI: CL0P turned MOVEit Transfer into a repeatable mass data-theft pathway via CVE-2023-34362
A joint CISA and FBI advisory details how the CL0P ransomware group (also tracked as TA505) exploited a previously unknown SQL injection flaw (CVE-2023-34362) in Progress Software’s MOVEit Transfer managed file transfer (MFT) product to implant a web shell (“LEMURLOOT”) and exfiltrate data from underlying databases. The advisory frames MOVEit as the latest example of a broader TA505 pattern: targeting internet-facing MFT platforms with zero-day exploits (Accellion FTA in 2020–2021, GoAnywhere MFT in early 2023, and MOVEit in May 2023) to conduct large-scale theft and extortion—often emphasizing data exfiltration over encryption.
Microsoft’s Secure Future Initiative: a multi-year, hyperscaler-scale reset on how Microsoft builds and operates security
Microsoft’s Secure Future Initiative (SFI), launched in November 2023, is a multi-year, cross-company program intended to “increasingly secure” how Microsoft designs, builds, tests, and operates its products and services. Microsoft says the first year prioritized security across the company through internal training and substantial engineering investment to reduce risk. SFI is structured around security principles (innovate, implement, guide) and six engineering pillars mapped to Zero Trust principles and the NIST Cybersecurity Framework, signaling a governance-and-engineering approach rather than a point-product response.
For global cloud, identity, and security teams, SFI matters because it describes Microsoft’s internal hardening focus areas—identity and secrets, tenant isolation, network segmentation, SDLC/build integrity, unified detection, and faster remediation—that can influence default configurations, platform controls, and operational expectations across Microsoft’s cloud and software ecosystem over time. Microsoft also publishes periodic SFI progress reports (including references to a November 2025 report and earlier updates), indicating the initiative is intended to be measured and iterated in “waves” as threats evolve.
HTTP/2 Rapid Reset (CVE-2023-44487) highlights a web-scale protocol abuse pattern for DDoS
Cloudflare documented a record-scale DDoS wave that abused HTTP/2 stream cancellation (RST_STREAM) to generate extreme request rates with a relatively small botnet. The “Rapid Reset” technique (tracked as CVE-2023-44487) exploits HTTP/2’s ability to open many concurrent streams and then instantly cancel them, letting attackers recycle concurrency slots faster than some servers and intermediaries can clean up state. Cloudflare said attacks began Aug. 25, 2023 and peaked just above 201 million requests per second, observed alongside similar activity reported by Google and AWS, prompting coordinated disclosure to vendors and critical infrastructure providers.
OWASP formalizes a shared security baseline for GenAI apps with the Top 10 for LLM Applications (now part of the broader GenAI Security Project)
OWASP’s Top 10 for Large Language Model (LLM) Applications has been published as a community security baseline that catalogs common failure modes in GenAI applications—ranging from prompt injection to model theft. OWASP says the effort has expanded beyond a list into the OWASP GenAI Security Project, a broader open initiative covering risks across LLMs, agentic systems, and AI-driven applications, with a large global contributor community and separate project resources and participation tracks.
Google’s SAIF reframed AI security as operational controls, not just model research
Google introduced the Secure AI Framework (SAIF) in June 2023 as a conceptual security framework for AI systems, explicitly mapping AI-specific threats (e.g., model theft, data poisoning, prompt injection, and training-data leakage) to familiar security disciplines such as secure-by-default infrastructure, detection and response, automation, consistent platform controls, continuous testing/feedback loops, and end-to-end risk assessment. While SAIF is not a standard, Google positioned it as a bridge between traditional security programs and emerging AI risks, and tied it to ongoing industry work including NIST’s AI Risk Management Framework and ISO/IEC 42001.
NIST AI RMF: the U.S. government’s voluntary baseline for AI trust, security, and resilience—now expanding to generative AI and critical infrastructure
NIST’s AI Risk Management Framework (AI RMF) established a shared, voluntary vocabulary and process model for managing AI risks across the lifecycle—supporting “trustworthiness” goals such as safety, security, and resilience. Since the AI RMF 1.0 release on Jan. 26, 2023, NIST has expanded implementation support via the AI RMF Playbook and Resource Center, published a Generative AI Profile (NIST-AI-600-1) in July 2024, and, as of Apr. 7, 2026, issued a concept note for a forthcoming profile focused on Trustworthy AI in Critical Infrastructure—signaling growing expectations that AI governance and security controls will be tailored to high-consequence environments.
CrowdStrike publishes RCA for July 2024 “Channel File 291” Windows sensor outage, reframing update resilience as a board-level risk
CrowdStrike released a root-cause analysis (RCA) and executive summary for the July 19, 2024 “Channel File 291” incident, in which a content configuration update delivered via channel files for its Windows sensor triggered a widespread outage. The company says the specific scenario is now incapable of recurring and outlines mitigations and process improvements intended to enhance resilience. CrowdStrike also reported that by July 29, 2024 at 8:00 p.m. EDT, approximately 99% of Windows sensors were back online, which it compares to a typical ~1% week-over-week variance in sensor connections.