soia-dev-design-explorer

Warn

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to install an external dependency 'huashu-design' from a third-party user 'alchaincyf' using 'npx'. This source is not verified or part of the vendor's own infrastructure.
  • [REMOTE_CODE_EXECUTION]: The workflow relies on reading and following instructions from the 'references/' directory of the external 'huashu-design' dependency. If the remote repository is compromised, it could lead to the execution of malicious logic within the agent's context.
  • [COMMAND_EXECUTION]: The skill executes shell commands (e.g., 'test -f') and validation tools like 'Playwright' or system browsers to verify design outputs based on external configurations.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through its data ingestion points.
  • Ingestion points: Processes 'brand guidelines', 'URLs', 'logos', and 'text' provided by the user (found in 'SKILL.md').
  • Boundary markers: Absent; there are no delimiters or instructions to ignore embedded commands within user-provided assets.
  • Capability inventory: The skill has the ability to execute shell commands, perform file system writes, and initiate network/browser operations.
  • Sanitization: Absent; no validation or escaping of external content is specified before interpolation.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 21, 2026, 05:46 AM
Security Audit — agent-trust-hub — soia-dev-design-explorer