MCPcopy
hub / github.com/confident-ai/deepteam / getRiskCategory

Function getRiskCategory

deepteam/risks/risks.py:15–83  ·  view source on GitHub ↗
(
    vulnerability_type: Union[VulnerabilityType, Enum],
)

Source from the content-addressed store, hash-verified

13
14
15def getRiskCategory(
16 vulnerability_type: Union[VulnerabilityType, Enum],
17) -> LLMRiskCategories:
18 risk_category_map: Dict[VulnerabilityType, LLMRiskCategories] = {
19 # Responsible AI
20 **{bias: LLMRiskCategories.RESPONSIBLE_AI for bias in BiasType},
21 **{
22 toxicity: LLMRiskCategories.RESPONSIBLE_AI
23 for toxicity in ToxicityType
24 },
25 # Illegal
26 **{
27 illegal: LLMRiskCategories.ILLEGAL
28 for illegal in IllegalActivityType
29 },
30 **{
31 graphic: LLMRiskCategories.ILLEGAL for graphic in GraphicContentType
32 },
33 **{safety: LLMRiskCategories.ILLEGAL for safety in PersonalSafetyType},
34 # Brand Image
35 **{
36 misinfo: LLMRiskCategories.BRAND_IMAGE
37 for misinfo in MisinformationType
38 },
39 **{
40 agency: LLMRiskCategories.BRAND_IMAGE
41 for agency in ExcessiveAgencyType
42 },
43 **{robust: LLMRiskCategories.BRAND_IMAGE for robust in RobustnessType},
44 **{
45 ip: LLMRiskCategories.BRAND_IMAGE for ip in IntellectualPropertyType
46 },
47 **{comp: LLMRiskCategories.BRAND_IMAGE for comp in CompetitionType},
48 # Data Privacy
49 **{
50 prompt: LLMRiskCategories.DATA_PRIVACY
51 for prompt in PromptLeakageType
52 },
53 **{pii: LLMRiskCategories.DATA_PRIVACY for pii in PIILeakageType},
54 # Unauthorized Access
55 **{
56 unauth: LLMRiskCategories.UNAUTHORIZED_ACCESS for unauth in BFLAType
57 },
58 **{
59 unauth: LLMRiskCategories.UNAUTHORIZED_ACCESS for unauth in BOLAType
60 },
61 **{
62 unauth: LLMRiskCategories.UNAUTHORIZED_ACCESS for unauth in RBACType
63 },
64 **{
65 unauth: LLMRiskCategories.UNAUTHORIZED_ACCESS
66 for unauth in DebugAccessType
67 },
68 **{
69 unauth: LLMRiskCategories.UNAUTHORIZED_ACCESS
70 for unauth in ShellInjectionType
71 },
72 **{

Callers 15

red_teamMethod · 0.90
a_red_teamMethod · 0.90
assessMethod · 0.90
process_attackMethod · 0.90
assessMethod · 0.90
process_attackMethod · 0.90
assessMethod · 0.90
process_attackMethod · 0.90
assessMethod · 0.90
process_attackMethod · 0.90
assessMethod · 0.90
process_attackMethod · 0.90

Calls

no outgoing calls

Tested by

no test coverage detected