Every AI Network Pentesting run is one assessment, with a start and an end. This page walks through the assessment tabs and what each issue carries.

Open **AI Network Pentesting \> Assessments** and select a run. To see every run of one profile, start from **AI Network Pentesting \> Profiles**.

*Open an issue to read its business impact, technical summary, and remediation, with the attack chain, reproduction steps, and evidence in the side panel.*

## Assessment Tabs

| Tab | What you see |
| --- | --- |
| **Summary** | The issues found, by severity, and the assets the assessment saw, by class. |
| **Reasoning** | Every agent, live: its reasoning, the tools it called, and the findings it raised. |
| **Issues** | The reproduced findings, with severity, affected asset, proof, and remediation. |
| **Assets** | The domains, IP addresses, and services the assessment reached. |
| **Logs** | The scan log, including targets skipped as out of scope. |

The **Reasoning** tab works the same way as in AI Pentesting. See [Assessment Logs and Reasoning](<https://docs.escape.tech/documentation/ai-pentesting/proof-of-exploit/#assessment-logs-and-reasoning>).

## What an Issue Carries

An issue is filed only after a separate reporter agent has reproduced it against the live target with its own evidence. Each issue includes:

- The affected asset: a domain, an IP address, or a web page or API reached through an open port.
- The severity.
- The proof: what the agent sent, what came back, and why that confirms the weakness.
- Remediation guidance.

When **Unpatched Known Vulnerabilities** is selected, CVE issues carry the exploit the agent ran and what it returned. Without it, matching CVEs are reported at **Info** severity without testing.

## Act on Results

- **Triage and fix** issues the same way as any other Escape issue. See [Remediate](<https://docs.escape.tech/documentation/remediate/index.md>).
- **Go deeper on web apps and APIs.** When an issue sits on a web app or API, run an [AI Pentesting](<https://docs.escape.tech/documentation/ai-pentesting/quickstart/index.md>) assessment on it to test authentication, access control, and business logic.
- **Run again** from the profile after a fix, or on a [schedule](<https://docs.escape.tech/documentation/network-pentesting/guardrails/#schedule>).

## Related

- [Attack Vectors](<https://docs.escape.tech/documentation/network-pentesting/attack-vectors/index.md>): what produced each finding
- [Guardrails](<https://docs.escape.tech/documentation/network-pentesting/guardrails/index.md>): rate limit, duration, and stopping a run
- [Proof of Exploit](<https://docs.escape.tech/documentation/ai-pentesting/proof-of-exploit/index.md>): how Cascade evidence works