#llm-security

Brakes for your AI agents: every tool call is checked against your rules, held for human approval when risky, and recorded in a tamper-evident audit trail. Works with Claude Code and any MCP app.

Security boundary CLI proxy between AI coding agents and your shell — redacts secrets, neutralizes prompt injection, wraps untrusted terminal output, and audits every command. Single Rust binary.

High-performance zero-dependency static analysis tool enforcing NASA Power of Ten rules across 20 programming languages, plus AI agent skill, MCP server config, prompt template, and LLM model security auditing.

Let AI agents query and update Postgres/MySQL without raw SQL, unrestricted schema access, or database credentials in the model; writes stay reviewed.

security

Python SDK for Silmaril Firewall prompt injection and jailbreak detection

No repository description published.

Know what you're letting in — scans, signs and verifies what an agent loads. Agent Skills vs the OWASP Agentic Skills Top 10, with SARIF output and a load-time gate.

TypeScript SDK for Declaw — security-first sandboxing for AI agents

A TypeScript library for validating and securing LLM prompts — PII, jailbreak, prompt injection, profanity, toxicity and relevance guards. Zero-config defaults, extensible with custom patterns.

Nation-state grade, defense-in-depth security testing framework. Tests web apps, extensions, desktop, mobile, cloud, AI/LLM, hardware, and supply chain from a single CLI.

25 production-tested defensive security skills for Claude Code - WordPress, VPS, Cloudflare, Next.js hardening, AI agent guardrails, MCP security, prompt injection defense, OWASP LLM Top 10, LLM coding failure modes (slopsquatting, hallucinated APIs, sycophancy), incident response, GDPR/DACH compliance. MIT, battle-tested.

A Postgres wire-protocol proxy that blocks AI SQL injection at the AST level. Protects databases from LLM agents (LangChain, LlamaIndex) by parsing and rewriting destructive queries on the fly.