> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getmuster.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Risk Scoring

> How Muster calculates risk scores for your AI agents.

Every agent in Muster gets a risk score from 0 to 100, calculated automatically from its attributes. The score determines the risk level shown in the dashboard.

## Risk levels

| Score  | Level    | Badge colour |
| ------ | -------- | ------------ |
| 70–100 | Critical | Red          |
| 50–69  | High     | Orange       |
| 25–49  | Medium   | Yellow       |
| 0–24   | Low      | Green        |

## Scoring rules

| Rule ID | Condition                            | Points |
| ------- | ------------------------------------ | ------ |
| R01     | Agent writes to external systems     | +20    |
| R02     | Agent handles PII data               | +15    |
| R03     | Agent makes external LLM API calls   | +15    |
| R05     | No owner email assigned              | +10    |
| R07     | Agent can execute code               | +12    |
| R14     | Framework detection confidence ≥ 85% | −5     |

The rules that fired for each agent are visible in the agent detail panel — fully transparent, no black box.

## Discrepancies

If an agent is registered as "no PII" but Muster's discovery scanners detect PII-related patterns in its code or API calls, the agent is flagged with a **discrepancy**. Discrepant agents appear in the inventory with a warning badge and are prioritised in the Health Heatmap.

## Risk score vs risk level

The score (0–100) is the raw value. The level (Critical/High/Medium/Low) is the band it falls into. Use the score for sorting and trend analysis; use the level for at-a-glance status.