excalidraw
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill recommends installing the 'cryptography' Python package. This is a standard and well-known library required for the script's encryption functionality.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a local Python script ('scripts/upload.py') through the terminal to share the generated diagrams. This action is the primary intended function of the skill.
- [INDIRECT_PROMPT_INJECTION]: The skill translates natural language descriptions into structural JSON diagrams, creating a surface for potential indirect prompt injection if the resulting diagram labels are later interpreted as instructions by an agent.
- Ingestion points: Data from the agent's context and user requests used to populate diagram labels and text elements (SKILL.md, references/examples.md).
- Boundary markers: None; there are no specific markers or instructions to delimit or ignore instructions within the text content of diagrams.
- Capability inventory: The skill can write files and execute a terminal script with network access ('scripts/upload.py').
- Sanitization: None; the skill does not mention validation or escaping for labels and text content within the JSON diagram structure.
Audit Metadata