excalidraw
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied .excalidraw and .excalidraw.json files, which creates a surface for indirect prompt injection. Malicious instructions could be embedded within the text elements of a diagram and then extracted by the subagent for processing by the main agent.
- Ingestion points: Operations are performed on files with .excalidraw or .excalidraw.json extensions in the local file system.
- Boundary markers: While the skill uses subagent delegation to isolate tasks, the provided templates do not include explicit safety delimiters or instructions to ignore potential commands embedded in the diagram's content.
- Capability inventory: The skill utilizes file reading and writing capabilities via the agent's task delegation tool.
- Sanitization: There is no evidence of sanitization or filtering of text labels extracted from the JSON files before the information is returned to the main agent context.
Audit Metadata