deco-core-architecture

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill documents a resolution engine that processes Resolvable objects (CMS state) from various DecofileProviders (filesystem, realtime CMS websocket, or JSON). This architecture allows external configuration to influence server-side logic and UI rendering. \n
  • Ingestion points: engine/decofile/ (filesystem, realtime, and JSON providers).\n
  • Boundary markers: The framework utilizes JSON Schema generation and TypeScript types for prop validation.\n
  • Capability inventory: Handlers can execute network fetch operations; Loaders and Actions perform server-side data mutations; Sections render Preact components.\n
  • Sanitization: Standard server-side rendering with Preact is used, though the useScript hook allows for raw script injection into the DOM.\n- [DYNAMIC_EXECUTION]: The documentation describes the useScript hook (hooks-components.md), which is used to dynamically minify and generate JavaScript code at runtime for client-side injection via dangerouslySetInnerHTML.\n- [COMMAND_EXECUTION]: The reference includes common development and production tasks defined in deno.json, such as start, dev, and gen, which involve executing framework-specific shell commands during the build and runtime phases.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 06:53 AM
Security Audit — agent-trust-hub — deco-core-architecture