skills/decentraland/sdk-skills/npcs/Gen Agent Trust Hub

npcs

Pass

Audited by Gen Agent Trust Hub on Oct 22, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing dcl-npc-toolkit using npm i. This is the official toolkit for creating NPCs in the Decentraland environment and is maintained by the vendor.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides mechanisms for processing dialogue text (npc.Dialog) and player metadata (e.g., wearables and names via getPlayer()). This represents a potential surface for indirect prompt injection where malicious instructions could be embedded in strings processed by the NPC's dialogue system.
  • Ingestion points: Dialogue definitions in SKILL.md and references/npc-library.mdc; Player profile data (names, wearables) retrieved via the getPlayer() API.
  • Boundary markers: The skill uses structured Dialog arrays to separate text from logic, but does not explicitly implement sanitization logic for user-provided strings.
  • Capability inventory: The skill is focused on UI rendering, animation, and movement within the Decentraland SDK; it does not include capabilities for arbitrary code execution or file system modifications based on the processed data.
  • Sanitization: No explicit sanitization or filtering of external dialogue content is described.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 22, 2026, 11:07 PM