keywords:llm-security

@claude-flow/aidefence

AI Manipulation Defense System (AIMDS) with self-learning, prompt injection detection, and vector search integration

published version 3.0.3, a month ago3 dependents licensed under $MIT

95,321

AI security skills grounded in mid-2026 threat reality, not stale framework documentation. 51 skills, 11 catalogs (459 CVEs / 181 CWEs / 805 ATT&CK + ICS / 170 ATLAS / 468 D3FEND / 8888 RFCs), 35 jurisdictions, 10-class catalog gap detector + budget gate,

dotcoocoo

published version 0.18.11, 6 hours ago0 dependents licensed under $Apache-2.0

36,660

@promptshield/sanitizer

PromptShield sanitizer that applies safe, deterministic fixes to text based on detected prompt-injection threats such as invisible characters, markdown smuggling, and BOM artifacts.

mayank1513

published version 1.0.0, 3 months ago4 dependents licensed under $MIT

23,836

@promptshield/ignore

Ignore directive parser and threat filtering utilities for PromptShield. Provides file-level and line-level suppression logic for CLI and IDE integrations.

mayank1513

published version 2.0.0, 3 months ago4 dependents licensed under $MIT

22,379

@mcp-guardian/server

Security, cost, and health governance proxy for MCP infrastructure — three-layer detection engine (regex + schema + LLM), monorepo, corpus, CI/CD

rudraneel93

published version 4.1.8, 15 days ago1 dependents licensed under $MIT

4,888

shellward

AI agent security & MCP security middleware — prompt injection detection, AI firewall, runtime guardrails & data-loss prevention for LLM tool calls. 8-layer defense against data exfiltration & dangerous commands. Zero dependencies. SDK + OpenClaw plugin.

jnmetacode

published version 0.7.16, 20 hours ago0 dependents licensed under $Apache-2.0

4,174

agent-threat-rules

Open detection standard -- like Sigma, but for AI agents. 655 rules for prompt injection, tool poisoning, context exfiltration, and MCP attacks. Shipped in Cisco AI Defense. 97.2% recall on NVIDIA garak.

panguard0414

published version 3.5.2, 14 hours ago5 dependents licensed under $MIT

3,304

ship-safe

AI-powered multi-agent security platform. 23 agents scan 80+ attack classes including AI integration supply chain (Vercel-class attacks), Hermes Agent deployments (ASI-01–ASI-10), tool registry poisoning, function-call injection, skill permission drift, a

asamassekou

published version 9.3.2, a month ago0 dependents licensed under $MIT

2,581

@silmaril-security/sdk

TypeScript SDK for Silmaril Firewall — prompt injection and jailbreak detection

GitHub Actions

published version 0.4.2, 20 days ago0 dependents licensed under $SEE LICENSE IN LICENSE

2,537

@sharkvoid/rasp

Production-grade Runtime Application Self-Protection for Express, Next.js, Firebase, and Supabase. Defends against SQL injection, XSS, path traversal, command injection, bots, brute force, and AI-powered attackers. Powered by SharkVoid.

sharkvoid_team

published version 2.2.8, 11 days ago1 dependents licensed under $MIT

2,342

zentric-protocol-mcp

Prompt injection detection + PII anonymization MCP server for AI agents, with a signed audit trail. Deterministic verdicts (CLEARED/ANONYMIZED/BLOCKED), 22 signatures, 7 languages. Every request returns a SHA-256-signed audit record — GDPR Art. 30 evidenc

zentricprotocol

published version 0.1.9, 12 days ago0 dependents licensed under $MIT

1,447

aidefence

AI Defense System — Production-ready security middleware for protecting AI applications from prompt injection, manipulation attacks, PII leakage, and adversarial inputs. Available as both 'aidefence' (British) and 'aidefense' (American).

ruvnet

published version 2.3.0, a month ago4 dependents licensed under $MIT OR Apache-2.0

1,322

agent-security-scanner-mcp

Security scanner MCP server for AI coding agents. Prompt injection firewall, package hallucination detection (4.3M+ packages), 1700+ vulnerability rules with AST & taint analysis, LLM-powered semantic code review, auto-fix. For Claude Code, Cursor, Windsu

prooflayer

published version 4.4.4, 2 hours ago0 dependents licensed under $MIT

1,211

@evalguard/cli

EvalGuard CLI - Run LLM evaluations and security scans from the command line

GitHub Actions

published version 3.3.0, 3 days ago1 dependents licensed under $Apache-2.0

978

ai-security-scan

AI-aware security scanner for source code, CI/CD pipelines and containers. Detects prompt injection, RAG/agent/MCP risks, secrets, and OWASP LLM Top 10 issues.

creatiaivity

published version 1.3.0, 3 days ago0 dependents licensed under $MIT

1,096

@evalguard/sdk

Official EvalGuard SDK — LLM evaluation, red-team security, runtime guardrails, observability, and FinOps. Note: a third-party package named `evalguard` exists on npm and is not affiliated with EvalGuard, Inc.

GitHub Actions

published version 2.4.0, 3 days ago3 dependents licensed under $Apache-2.0

871

@evalguard/core

EvalGuard core — LLM evaluation, security testing, firewall, gateway, and monitoring engine. Runtime dep of @evalguard/sdk and @evalguard/cli.

GitHub Actions

published version 1.4.0, 3 days ago2 dependents licensed under $Apache-2.0

880

security-mcp

AI security MCP server and enforcement gate for Claude Code, Cursor, GitHub Copilot, Codex, Replit, and any MCP-compatible editor. Applies OWASP, MITRE ATT&CK, NIST, Zero Trust, PCI DSS, SOC 2, and ISO 27001.