> ## Documentation Index
> Fetch the complete documentation index at: https://docs.reclaimllm.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Detect Risky AI Coding Sessions with Signals
> Use ReclaimLLM Signals to find where AI coding issues concentrate, triage anomaly reports by severity, and understand each anomaly code.
When your team codes with AI assistants, problems hide in hundreds of sessions. Signals shows you where to look: the overview for the last 30 days walks you from where issues concentrate, to who owns them, to the evidence sessions behind them.
## Triage signals
Start on the Signals overview. The **Projects × patterns** matrix shows which projects keep hitting the same problem, such as long idle gaps in sessions.
In **Anomaly reports**, filter by status, severity, or detector. Filter to **High** severity to triage what matters first.
Open any code in **What the anomaly codes mean** (AD-001 to AD-012). Each explains what it looks like, what an admin should do, and the guardrail to set.
Click a matrix cell to jump to the evidence sessions behind it.
Acknowledge, Resolve, Dismiss, or Accept risk on each report.
## FAQ
AD-003 is a runaway or looping session. The code detail explains what it looks like, the admin action, and the guardrail to set.
Filter Anomaly reports to High severity, then work through them with Acknowledge, Resolve, Dismiss, or Accept risk.
Yes. Click any cell in the Projects × patterns matrix to open the evidence sessions behind it.
## Related
* Reference: [Signals tab](/enterprise/dashboard/signals)
* Product: [Enterprise observability](https://reclaimllm.com/features/enterprise-observability)
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.
## FAQ