skills/rsamaium/rpg-js/rpgjs/Gen Agent Trust Hub

rpgjs

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches documentation indices and guides from project-specific domains (v5.rpgjs.dev, canvasengine.net) and the official RPGJS GitHub repository.
  • [COMMAND_EXECUTION]: Instructs the agent to use curl for downloading remote files and ripgrep (rg) for text extraction. It also describes standard project setup procedures using package managers.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external text from documentation sources, which presents a surface for indirect prompt injection. This is mitigated by the use of official vendor documentation as the source.
  • Ingestion points: Documentation files retrieved from v5.rpgjs.dev and canvasengine.net (e.g., llms.txt).
  • Boundary markers: No explicit delimiters are present in the instructions to separate documentation content from system prompts.
  • Capability inventory: File system access, network access (curl), and shell command execution (ripgrep, npm/yarn).
  • Sanitization: Content is ingested as raw text without validation or sanitization layers.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 02:04 PM
Security Audit — agent-trust-hub — rpgjs