ui-consolidation
Pass
Audited by Gen Agent Trust Hub on Jul 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands such as
grepand the repository's type-check command to verify the codebase after component consolidation.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes source code from various project directories and has the capability to modify or delete files and execute shell commands based on the analyzed content.\n - Ingestion points: Source code files located in
<design-system>/components/,<app>/components/shared/, and<app>/routes/**/(SKILL.md).\n - Boundary markers: None identified; the skill does not specify markers to delimit the code content or instructions to ignore embedded prompts.\n
- Capability inventory: Includes file reading (discovery), file writing/modification (refactoring), file deletion (cleanup), and shell command execution (grep, typecheck) (SKILL.md).\n
- Sanitization: No specific sanitization or validation steps are defined for the ingested source code content.
Audit Metadata