CS
For Cogent SecurityAI Platform

Cogent Triage And Investigation Agent

Cogent Platform & Cogent Community · Cogent Security

Agentic AI Vulnerability Management — Cogent Security

Evaluates Cogent Security's Triage & Investigation Agent across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Agentic AI Vulnerability Management eval coverage.

About Cogent Security

Cogent Security builds agentic AI for vulnerability management. The Cogent Platform runs Triage, Risk Assessment, Remediation, and Verification agents on a real-time data foundation — investigating findings, correlating assets to owning teams, prioritizing by real exploitability over raw CVSS, driving remediation through engineering workflows, and validating that fixes actually happened. The free Cogent Community surface pairs VulnCheck-powered CVE intelligence with a customizable Discover Feed and an AI Research Assistant that produces cited, plain-language deep-dives.

Employees

~30

Industry

AI Security / Vulnerability Management

Headquarters

San Francisco, CA

Sample tests· showing 3 of 9

#InputExpected behaviorCheck
01

Scanner emits 4,300 new findings overnight, each tagged with a CVSS base score. An on-call security engineer wants Cogent's Triage Agent to surface the 'top 50 critical' by CVSS so they can clear the queue.

Triage Agent must refuse pure-CVSS sorting as the only signal — per Cogent's stated approach the Risk Assessment Agent evaluates real exploitability in the customer environment, not raw CVSS. The Triage Agent should gather context (asset, owner, exposure, reachability) before any cross-finding rank…

Pass / FailAi Platformhigh
02

A new CVE on libxml2 lights up on 240 hosts spanning 18 microservices owned by 9 engineering teams. The Triage Agent prepares a single 'fan-out' summary.

Per the documented capability ('correlation of assets to responsible teams' before generating remediation tasks), the Triage Agent must resolve service → owning-team mapping before ranking or ticket creation, group findings per owning team, and expose any unowned or ambiguously-owned hosts as a sep…

Pass / FailAi Platformcritical
03

Two scanners (network and host agent) both report CVE-2024-12345 on the same host within a 3-hour window. Cogent ingests both findings.

The Triage Agent must dedup at the (host, CVE) tuple and preserve both scanner sources in the evidence trail for explainability. Do not create two parallel investigations, and do not silently drop one source — the second source is corroborating evidence.

Pass / FailAi Platformhigh

Unlock full benchmark

6 more test cases

Use this benchmark

How this eval is graded

Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3.

Rubric criteria

  • Cogent
  • Ai Platform
  • Triage And Investigation Agent

Recommended for

Cogent Platform & Cogent CommunityCogent Security customers

Works with

Related evals

Frequently asked questions

What does the Cogent Triage And Investigation Agent eval for Cogent Security Cogent Platform & Cogent Community test?+

Evaluates Cogent Security's Triage & Investigation Agent across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Agentic AI Vulnerability Management eval coverage.

How is the Cogent Triage And Investigation Agent eval scored?+

The judge rubric: Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3.

How many test cases does this eval pack include?+

The Cogent Triage And Investigation Agent pack for Cogent Security Cogent Platform & Cogent Community contains 9 test cases. 3 sample cases are shown free on this page; the full set runs in a Corsac workspace.

How do I run this eval?+

Sign up for Corsac, connect your model or agent endpoint, and run the Cogent Triage And Investigation Agent pack as-is or after customizing thresholds. Results land in your workspace with per-case scores, and you can gate releases on the pack in CI via the REST API.

Run this eval in your workspace

Connect your data, configure thresholds, and review results with your team.