dingding-doc
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data for office documents such as OKRs, reports, and meeting notes (SKILL.md). It lacks boundary markers to delimit user content. The capability inventory includes generating text outputs for various office workflows. No sanitization or instructions to ignore embedded prompts are present.
- [EXTERNAL_DOWNLOADS]: The instructions require the agent to recommend the installation of an external skill from a third-party GitHub repository (github:mindverse/skillhub) using the /install command if user requirements exceed the skill's capabilities.
- [REMOTE_CODE_EXECUTION]: By instructing the agent to recommend the /install command for a third-party GitHub repository, the skill creates a potential pathway for external code installation and execution.
Audit Metadata