Skill UI

Authoring and tuning Falco rules to detect container escape techniques, including host filesystem mounts, nsenter usage, privileged containers, kernel module loading, and host manipulation. Covers rule syntax, condition tuning, output fields, priorities, and false positive reduction.

This skill automates the Design Space Exploration (DSE) process, crucial for computer architecture and Electronic Design Automation (EDA). It runs complex simulations, analyzes performance metrics, and systematically tunes parameters (e.g., cache size, pipeline width) across a defined parameter space. It iterates until the objective metric is optimized or a defined timeout is reached, making it ideal for complex hardware tuning and system design.

This skill provides a structured methodology for mitigating alert fatigue within Security Operations Centers (SOCs). It guides users through advanced techniques, including implementing Risk-Based Alerting (RBA) to replace high-volume, low-fidelity alerts with consolidated risk scores. It emphasizes quantifying alert quality by analyzing True Positive (TP) and False Positive (FP) rates, and systematically tuning detection rules to maintain analyst effectiveness.

This skill covers the deployment and fine-tuning of Web Application Firewall (WAF) rules across major cloud platforms like AWS, Azure, and Cloudflare. It addresses critical security needs such as mitigating OWASP vulnerabilities (SQLi, XSS), managing bot traffic, implementing rate limiting, and ensuring compliance (e.g., PCI-DSS). The workflow guides users through deploying managed rule sets, creating custom rate-based logic, and minimizing false positives.

This skill guides security teams on systematically tuning SIEM detection rules in platforms like Splunk and Elastic. It focuses on reducing alert fatigue and false positives by analyzing historical alert volumes, establishing environmental baselines, creating context-aware whitelists, and measuring detection efficacy using metrics like precision and recall. This is crucial for maturing a Security Operations Center (SOC).

This guide details configuring ModSecurity WAF and OWASP Core Rule Set (CRS) for comprehensive web application security auditing. It covers essential steps such as initial deployment, rule tuning to minimize false positives, configuring detailed audit logging, and integrating logs into SIEM systems for robust threat detection.

This guide details how to integrate Static Application Security Testing (SAST) tools like CodeQL and Semgrep into your CI/CD pipelines using GitHub Actions. It covers setting up automated vulnerability detection on pull requests, tuning rules to minimize false positives, uploading SARIF results for centralized security reporting, and enforcing quality gates to prevent vulnerable code from merging into protected branches. Essential for DevSecOps and compliance (SOC 2, PCI DSS).

A comprehensive guide for Security Operations Center (SOC) analysts on reducing false positive alerts in SIEM systems. Techniques include systematic threshold tuning, refining correlation logic, implementing allowlisting, and enriching data with threat intelligence to combat alert fatigue and improve detection accuracy.

Guides fine-tuning and evaluation of OpenVLA-OFT/OFT+ for robot action generation, covering LoRA adaptation, continuous action heads, FiLM conditioning, and LIBERO/ALOHA benchmarks plus deployment considerations.

Provides end-to-end guidance for fine-tuning and serving Physical Intelligence’s OpenPI pi0 family (pi0, pi0-fast, pi0.5) with both JAX and PyTorch backends, covering dataset prep, norm stats, checkpoint conversion, and low-latency WebSocket policy inference for ALOHA, DROID, and LIBERO tasks.

Comprehensive Snowflake development assistant covering SQL best practices, dynamic tables, streams, tasks, Snowpipe, Cortex AI functions/agents, Snowpark Python, dbt integration, performance tuning, and security hardening for production data platforms.

This skill provides a managed cloud environment for training and fine-tuning advanced vision models. Users can build and customize models for object detection (e.g., DETR, YOLOS), image classification, and segmentation (SAM/SAM2). By leveraging Hugging Face Jobs, training is executed on powerful cloud GPUs without requiring local GPU setup, and all resulting models are automatically saved to the Hugging Face Hub.