hermesoffice-decks
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill instructs the agent to use
execute_slide_scriptfor handling complex, multi-element layout changes. This tool enables the generation and execution of scripts at runtime based on real-time slide geometry and content. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from untrusted sources, including user briefs and
web_searchresults used for charts. This data is then used to populate slide elements and may influence the logic of generated layout scripts. - Ingestion points: The skill processes user-provided content for slide planning and external data retrieved via the
web_searchtool (SKILL.md). - Boundary markers: There are no instructions specifying the use of delimiters or warnings to ignore embedded instructions within the processed data.
- Capability inventory: The skill utilizes
execute_slide_scriptfor dynamic execution,web_searchfor data retrieval, and variousset_elementtools for content injection. - Sanitization: The instructions do not define any sanitization, validation, or escaping procedures for external content before it is processed or interpolated.
Audit Metadata