worldos-war-authoring
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill includes explicit security and integrity instructions, directing the agent never to bypass access controls through unauthorized means (such as SQL injection, web scraping, or private APIs) and ensuring creator-facing internal prompts and configurations are not exposed to players.
- [COMMAND_EXECUTION]: Interacts with the WorldOS platform through a set of Model Context Protocol (MCP) tools for fetching authoring guides, templates, and application configurations.
- [PROMPT_INJECTION]: The skill processes external data by remixing existing worlds and using templates to construct prompts. Ingestion points include 'get_authoring_template' and 'get_remixable_world_prompt' in SKILL.md. While specific boundary markers are not defined, the skill mitigates risk by requiring the use of authoritative platform templates and a mandatory validation step ('validate_world') for the final assembly.
Audit Metadata