sketch-design-from-reference
Pass
Audited by Gen Agent Trust Hub on Jun 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill makes extensive use of the
run_codetool to execute dynamically generated JavaScript against the Sketch internal API. This is used for creating frames, setting styles, and managing symbols. - [EXTERNAL_DOWNLOADS]: The agent is instructed to fetch data from external URLs provided by the user to extract DOM structures, CSS variables, and asset links for use in the design process.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it processes untrusted data (HTML, source code) from external websites. This content could contain instructions designed to influence the agent's output.
- Ingestion points: Step 1 in
SKILL.mddetails the extraction of content and source code from external URLs. - Boundary markers: There are no explicit instructions to use delimiters or ignore instructions found within the reference content.
- Capability inventory: The skill utilizes
run_codefor structural document changes andget_screenshotfor fidelity validation. - Sanitization: The instructions do not include steps for sanitizing or filtering text or code extracted from external references.
Audit Metadata