litestar-autowire

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill describes the AutowireLoader feature, which dynamically resolves and executes Python callables from dotted strings provided in the configuration (e.g., 'pkg.module:callable').- [REMOTE_CODE_EXECUTION]: The core discovery mechanism automatically imports Python modules from filesystem paths defined in the domain_packages configuration during application startup.- [PROMPT_INJECTION]: The automatic module discovery and integration loading create an attack surface for indirect prompt injection if the project structure or configuration strings are influenced by untrusted data.
  • Ingestion points: The domain_packages and integrations configuration fields in SKILL.md.
  • Boundary markers: None present for the imported code content.
  • Capability inventory: Includes module traversal using pkgutil, dynamic module importing, and execution of controller and listener logic.
  • Sanitization: The plugin performs validation to ensure that discovered objects are subclasses of Litestar's Controller or instances of EventListener.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 10:54 AM
Security Audit — agent-trust-hub — litestar-autowire