Tag

Runs AI-generated code in secure, isolated cloud sandboxes you control via Python or JavaScript SDKs; supports self-hosting (Terraform) and AWS/GCP, enabling agents and code-interpreting workflows to execute real-world tools safely.

CLI and library for automated LLM prompt and red-team evaluations — run local-sensitive evals, compare models, and integrate checks into CI/CD with declarative configs and built-in red‑teaming.

Scans LLMs for security and safety failures — probing for hallucination, data leakage, prompt injection, jailbreaks, toxicity, and misinformation. A CLI red‑teaming kit that runs static, dynamic and adaptive probes across many providers and outputs structured JSONL reports.

Runs a minimal, sandboxed subset of Python inside a Rust interpreter so LLM agents can safely execute generated code with microsecond startup and tight host-control. Blocks filesystem/env/network by default, supports external host calls, snapshots, resource limits, and bindings for Python, Rust, and JavaScript.

Detects file content types with a compact deep‑learning model that runs in milliseconds on a single CPU. Trained on ~100M samples across 200+ content types; offered as a Rust CLI plus Python, JS, and Go bindings for large‑scale security and file‑routing use.

Provides a modular Python framework to run standardized evaluations of large language models — including prompt engineering, tool usage, multi-turn dialog and model-graded scoring. Ships with 100+ pre-built evaluations and extension points for custom elicitation and scoring; intended for model comparison, safety checks and benchmark automation.

Packs a Git repository into a single AI-friendly file for easy ingestion by LLMs. Offers per-file and total token counts, optional Tree-sitter compression, secret scanning, and multiple interfaces (CLI, web, browser extension, Docker, MCP) for AI-driven code review and analysis.

Runs autonomous LLM-driven penetration-testing agents that discover, exploit, and produce reproducible vulnerability reports in sandboxed Docker environments. Ships with 20+ pentest tools, a knowledge-graph memory, multi-LLM provider support, and REST/GraphQL APIs for self-hosted deployments.

Exposes multiple databases to MCP-compatible LLM clients via a zero-dependency, token-efficient MCP server. Ships execute_sql/search_objects/custom-tools, a web workbench, and safety guardrails (read-only, row limits, timeouts).

Exposes a local MCP server that lets LLMs (e.g., Claude Desktop) query decompiled Android app context from a modified JADX GUI—supporting class/method retrieval, resources, xrefs, and debugger hooks for interactive reverse engineering workflows.

Converts commodity WiFi CSI into real-time DensePose maps, breathing/heart-rate, and presence detection without cameras; runs on an ESP32 sensor mesh and a Rust sensing server with Docker images, designed for edge-first, privacy-focused deployments.

A code-first collection of runnable tutorials for building production-ready generative-AI agents — step-by-step guides covering stateful workflows, vector memory, RAG, tool integrations, Docker/AWS/RunPod deployment, security guardrails, observability, and multi-agent patterns.