create-canvas
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from repository files to construct the canvas UI and has the capability to modify project source files, representing a potential indirect prompt injection surface.
- Ingestion points: Reads repository rules, routes, component APIs, themes, design tokens, and existing canvas files.
- Boundary markers: None explicitly mentioned in instructions.
- Capability inventory: Modifies project source files, executes repository-required shell checks, and accesses the clipboard API.
- Sanitization: Includes specific prohibitions against evaluating stored strings or model responses as code at runtime.
- [COMMAND_EXECUTION]: The skill requires the agent to run repository-required checks, which involves executing shell commands to validate imports and route mounting.
- [SAFE]: The skill explicitly forbids the use of production credentials or personal data, requires isolation for component previews, and warns against patching browser APIs or libraries.
Audit Metadata