Results for “malware”

Detect poisoned training data and backdoored models across the ML pipeline using statistical analysis, activation clustering, and spectral signatures.

Supply-chain malware infection scanner. IoC-based local scan + safe eradication for npm/PyPI worm campaigns (Mini Shai-Hulud 1st/2nd, S1ngularity, lottie-player). Detects OS persistence (LaunchAgent/systemd), IDE-hook implants (.claude/.vscode/.github/workflows), lockfile-pinned malicious versions, and known C2/Session-Protocol exfil traces. Orchestrates credential rotation in the correct order so revocation does not trigger the `rm -rf ~/` retaliation payload. Don't use for static SAST (Sentinel), skill/MCP/`.claude/` supply-chain audit (Chain), Sigma/YARA rule authoring (Vigil), or incident coordination (Triage).

Add programmable safety guardrails to LLM applications at runtime, including jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, and toxicity detection.

NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, toxicity detection. Uses Colang 2.0 DSL for programmable rails. Production-ready, runs on T4 GPU.

Extract embedded configuration from Agent Tesla RAT samples including SMTP/FTP/Telegram exfiltration credentials, keylogger settings, and C2 endpoints using .NET decompilation and memory analysis.