CrowdStrike Agent Graph Shows How Agents Act on the Endpoint
Blog post from Crowdstrike
CrowdStrike describes Agent Graph as the relational foundation of its Falcon Guardian product for monitoring and securing AI agents on endpoints at runtime. Built on the existing Falcon sensor, it links prompts, users, hosts, models, tools, MCP servers, processes, files, network connections, and detections into causal records of agent activity. Falcon Guardian uses this information for an agent inventory that identifies sanctioned and shadow agents, session graphs that trace activity from prompt to resulting system actions, and threat scores intended to prioritize risk investigation. For supported tools including Claude Code, OpenAI Codex, GitHub Copilot, OpenCode, and Cursor, agent hooks can capture prompt, tool-call, and output details and apply policies to block risky actions before execution.
| Trend | Post Mentions | Total Month Mentions | Posts | Companies | MoM |
|---|---|---|---|---|---|
| AI Agents | 22 | 6,890 | 1,628 | 266 | +23% |
| AI Guardrails | 9 | 552 | 205 | 59 | +22% |
| Real-time | 6 | 4,128 | 1,085 | 233 | -15% |
| MCP | 3 | 11,696 | 1,199 | 231 | +10% |
| AI Coding Assistant | 1 | 1,929 | 531 | 146 | +21% |
| LLM | 1 | 6,496 | 1,223 | 229 | +8% |
| Local AI | 1 | 168 | 69 | 20 | -22% |
| Observability | 1 | 3,023 | 723 | 183 | +9% |
Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.