Skip to content

Threat Analyzer Agent

The [threat_analyzer_agent] section controls how the AI agent evaluates threats and mitigations for your repository.

Configuration reference

[threat_analyzer_agent]
litellm_provider = "azure-ai"
max_token_per_analysis = 8000
extra_context = []
excluded_threats = []

litellm_provider

Type: string — Required

Name of the LLM provider to use for this agent. Must match a key defined in [litellm_provider.*].

litellm_provider = "azure-ai"

max_token_per_analysis

Type: integer — Default: 8000

Maximum number of tokens the agent may use per analysis run.

max_token_per_analysis = 16000

number_of_concurrent_analysis

Type: integer — Default: 4

Number of concurrent analysis tasks the agent may run in parallel. Adjust based on your system resources and provider rate limits.

number_of_concurrent_analysis = 8

extra_context

Type: list[string] (file paths)

Paths to additional files whose content is appended to the agent's system prompt. Provide threat modeling context, prior assessments, or architecture documents to improve results.

extra_context = [
    "docs/threat-model-notes.md",
]

excluded_threats

Type: list[string]

Specific threat IDs to exclude from analysis. Use this to skip threats that are known to be irrelevant.

excluded_threats = ["TID-110", "TID-222"]

Example

[threat_analyzer_agent]
litellm_provider = "gh-copilot-sonnet"
max_token_per_analysis = 10000
extra_context = ["docs/security-assessment.md"]
excluded_threats = ["TID-110"]