Greenlight builds Luna, a security operations agent on AgentCore: up to 4,000 cases a month, more than 90% resolved without human intervention
Real numbers from a two-tier multi-agent security operations system: autonomous resolution, escalation rates and how human review is tiered by risk.
// Key points
- Roughly 3,000 to 4,000 alerts a month from four sources: dark web breach cards, identity and access anomalies, infrastructure alerts and financial risk events
- A Tier 1 supervisor orchestrates more than 30 specialized agents for triage; about 10% of cases needing multi-step reasoning go to a Tier 2 supervisor that correlates 3 to 5 data sources into an investigation report
- Autonomous resolution went from 0% to 83% with Tier 1 and more than 90% with both tiers, escalations fell below 10%, routine triage dropped from 15–30 minutes to seconds, and AWS estimates 900 to 1,800 analyst hours avoided per month
- A 30-day rolling episodic memory spots recurring patterns; a pattern counts as confirmed only after it is resolved the same way three times; high-risk cases go to humans with the full evidence chain
Builder's takeThe thing to copy is not the 90% but two rules: a pattern counts only after three identical resolutions, and risk tiers decide who signs off. I would apply the same to anomaly review in our interview product: low risk passes automatically, medium gets flagged, high goes to a person with the evidence attached.
// Background · from #AI agents
Full timeline →- Nov 9 Google measures agent distraction: 13% of sessions meet bug-referencing TODOs, and detoured sessions take 70% more steps and 77% more tokens
- Nov 9 Okta Cross App Access on AgentCore, part 2: all traffic through one Gateway, Cedar policies in default-deny, and denied calls never redeem a token
- Nov 9 AWS publishes a reference architecture for a data lake MCP server: Fargate in private subnets, SELECT-only Athena tools, and one JSON metadata file to onboard a dataset