excalidraw-diagram
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided text content to generate visual diagrams and automatically writes the output to the filesystem. It lacks robust boundary markers or sanitization to prevent malicious instructions embedded in the input from being written into the generated files or influencing the filename.
- Ingestion points: Processes arbitrary user-provided text content to identify concepts, relationships, and titles as described in the Workflow and Implementation Notes of SKILL.md.
- Boundary markers: The skill does not define explicit delimiters or instructions to ignore embedded commands when the agent processes the user input.
- Capability inventory: The skill instructs the agent to use a file-writing tool to automatically create and save .md files to the current working directory, which represents a capability surface that can be influenced by input.
- Sanitization: While the skill specifies replacing double quotes and parentheses for JSON compatibility, it lacks sanitization for potential path traversal in the generated filenames or for malicious markdown/plugin instructions within the diagram data.
Audit Metadata