od-deck-guizang-ppt
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill directly interpolates user-provided input ($ARGUMENTS) into a design brief used for tool execution and file generation.
- Ingestion points: The $ARGUMENTS variable in SKILL.md captures user input which is then embedded into a multi-step instruction set.
- Boundary markers: There are no explicit delimiters or boundary markers (such as XML tags or clear separators) to isolate user-provided content from the skill's own instructions within the generated brief.
- Capability inventory: The agent is authorized to call MCP tools (prepare_open_design_brief, register_open_design_artifact) and perform file-editing operations on the local system based on instructions returned by those tools.
- Sanitization: No sanitization, validation, or escaping of the $ARGUMENTS content is performed before it is passed to the next stage of the workflow.
- [DYNAMIC_EXECUTION]: The skill instructs the agent to author files based on instructions dynamically returned by a tool call (prepare_open_design_brief). This allows runtime-generated content from the tool's output to drive local file system modifications.
Audit Metadata