GenPark AI Agent Skill - Resource governor enforcing strict cgroup v2 CPU bandwidth quotas, wall-clock timeouts and memory OOM tripwires.
GenPark AI Agent Skill - Archive decompression security validator neutralizing Zip Slip and Tar Slip path traversal escapes.
GenPark AI Agent Skill - In-memory virtual filesystem with copy-on-write (COW) branching, revision rollback and snapshot isolation.
GenPark AI Agent Skill - Generates least-privilege Linux Seccomp BPF syscall filter profiles and blocks prohibited socket/ptrace operations.
GenPark AI Agent Skill - In-memory virtual filesystem with copy-on-write (COW) branching, revision rollback and snapshot isolation.
GenPark AI Agent Skill - Zero-dependency IP/CIDR egress packet validator restricting agent outbound traffic to trusted registries.
GenPark AI Agent Skill - Generates least-privilege Linux Seccomp BPF syscall filter profiles and blocks prohibited socket/ptrace operations.
GenPark AI Agent Skill - Archive decompression security validator neutralizing Zip Slip and Tar Slip path traversal escapes.
GenPark AI Agent Skill - Zero-dependency IP/CIDR egress packet validator restricting agent outbound traffic to trusted registries.
GenPark AI Agent Skill - Resource governor enforcing strict cgroup v2 CPU bandwidth quotas, wall-clock timeouts and memory OOM tripwires.
A memory and continuous learning system designed exclusively for AI — your AI grows smarter the more you use it.
A highly efficient, collaborative agentic coding framework using Model Context Protocol (MCP) to divide work between planning, code graph analysis, and execution.
Official Model Context Protocol (MCP) server for local-first PDF/A generation. Empowers AI agents (Claude, Cursor, Zed) to securely create, sign (PAdES), verify, and transform documents with barcodes and AcroForms.
This Python application scans for the CVE-2023-38831 vulnerability in WinRAR.
agent-sandbox ai-agents windows sandbox mxc microsoft-execution-container processcontainer typescript nodejs cursor claude-code agent-sandboxing wsl2 npm sdk security blast-radius
Governed runtime for autonomous AI agents with capability permissions, policy enforcement, and sandboxed execution.
Quantum Field Theory security framework using Feynman diagrams for dynamic ISO management. Implements physics-based isolation where attacks cause decoherence, not direct compromise. Three-state architecture: Closed (Ụzọ Mechiri), Open (Ụzọ Mepere), Protected (Ụzọ Nchekwa). When quantum systems fail, build your own.
A security engineer's playbook for defending agentic AI coding assistants (Claude Code, Cowork, Codex, Cursor) across laptops, remote workspaces, and GPU/sandbox clouds — structured as an LLM Wiki.
To associate your repository with the sandbox-security topic, visit your repo's landing page and select "manage topics."