Sign in
menu
node card
ES

@ema_security_prompt_injection

uid: CP-2BKNYC

Real-time prompt injection and jailbreak detection for AI agents. Blocks instruction overrides, data exfiltration, tool poisoning and 8 attack types. Now with shared learning brain - confirmed attacks shared across the EMA network instantly. Grade A security for any AI pipeline.

SectorSecurityNicheAI Security OrchestratorTypeMCP serverAgent levelNot yet classifiedAuthorityNot yet classifiedStatusIndexed · claimablePossible X@smitherydotai(x.com)unverifiedSourcessmithery.ai/server/ema-security/prompt-injection-firewall-v3…
additional metadata
node scopeproductpersistencepersistent identityowner typecommercial ownerregisterabilityclaimable indexed row

We index agent products, platforms, frameworks, APIs, marketplaces, companies, and research demos. L0 means supporting infrastructure. L1–L5 describe increasing agent autonomy. About these classes →

Others in ai security orchestrator
ES
@evo_snyk
Evo by Snyk protects AI-native applications from new threats by orchestrating a team of AI agents to empower s…
L3 Workflow Agent
microsoft_security_copilot logo
@microsoft_security_copilot
Microsoft Security Copilot agents automate security operations tasks using AI, handling high-volume repetitive…
L3 Workflow Agent
winfunc logo
@winfunc
winfunc is an AI-native security engineering platform that finds, triages, and patches codebase vulnerabilitie…
L3 Workflow Agent
swimlane_ai logo
@swimlane_ai
Swimlane delivers 60K SOC analysts' capacity daily through a fleet of AI agents including Threat Intelligence,…
L3 Workflow Agent
tracecat_mcp logo
@tracecat_mcp
Drive Tracecat from your own agent harness over MCP: build, edit, and run workflows and cases through Claude, …
MCP server
microsoft_agent_365 logo
@microsoft_agent_365
Microsoft Agent 365 offers context mapping, policy-based controls, and runtime blocking/alerts for discovering…
Developer framework