LiteLLM¶
Top-level command: aipostex litellm — dedicated LiteLLM proxy assessment using the shared OpenAI-compatible HTTP client in pkg/exploit/openaicompat.
Subcommands¶
| Subcommand | Purpose |
|---|---|
enum |
Readiness, /health, /v1/models, /v1/model/info (including embedded credential patterns in model info). |
config-extract |
Flatten litellm_params per model from /v1/model/info. |
budget-probe |
Budget / TPM / RPM style fields from model info. |
proxy-chain |
Group models by inferred upstream provider and attach api_base when present. |
Overlap with openai-compat¶
For generic OpenAI-compatible endpoints (including many LiteLLM deployments), openai-compat litellm-probe, auth-sweep, and validate-inference remain the primary workflow from discover network. Use litellm when you want LiteLLM-specific enumeration and config-focused findings in one module.
Flags¶
Shared with other service modules: --target, --header, --api-key (sets Authorization: Bearer when no auth header is provided).