component-diagram
Pass
Audited by Gen Agent Trust Hub on Jul 26, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it ingests user-provided prose to generate diagram models.
- Ingestion points: The skill parses user input to build a normalized diagram model as defined in the SKILL.md workflow.
- Boundary markers: No explicit delimiters are used to wrap or sanitize the user input within the agent's context.
- Capability inventory: The skill can execute JavaScript in the tldraw renderer and manage files via shell scripts.
- Sanitization: A parsing step is performed to normalize the input, providing a basic layer of validation before command generation.
- [COMMAND_EXECUTION]: The skill utilizes local shell scripts (scripts/tq and scripts/new-canvas) to interact with the tldraw application. These scripts access local application configuration files to discover the necessary API port and authorization token.
- [REMOTE_CODE_EXECUTION]: The skill dynamically generates JavaScript code and sends it to the tldraw application's local execution endpoint. This is the intended functional mechanism for updating the canvas in the offline application and is confined to the local host.
Audit Metadata