Threat intelligence for the age of AI agents.
AI agents now browse, code, pay, and talk to each other — and attackers have followed them there. ThreatPulse tracks the threats aimed at agents and the pathways between them, explains each one in plain English, and tells you what to do about it. Conventional vulnerabilities included, one click away.
Agent threats first
Prompt injection, MCP tool poisoning, memory poisoning, agent worms, and framework CVEs — classified against the OWASP Agentic Top 10 with attack-surface and propagation tags.
Plain English, written by AI
Claude turns every raw advisory into a readable threat profile with severity and remediation steps. Gemini writes a daily briefing of what actually matters this week.
From alert to action
On-demand incident-response playbooks for every threat, plus a chat advisor that specializes in securing AI agent deployments. Conventional CVEs stay on the watchlist too.
Explore the threat catalog
AI agent threats
Every threat to agents and agent-to-agent systems, newest first, classified by attack type.
OWASP Agentic Top 10, explained
ASI01–ASI10 in plain English, each with live examples from the feed.
Conventional watchlist
Actively exploited CVEs, zero-days, ransomware, and APT activity.
Top attack types: Prompt injection · MCP tool poisoning · Memory poisoning · Agent worms · Unsafe code execution
Continuously monitoring
Deduplicated at ingestion, synthesized by Claude, briefed daily by Gemini — free to start.