prism-ui
Pass
Audited by Gen Agent Trust Hub on Jul 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
bashtool to execute a Python module command:python3 -m prism_agent.uiux search --domain react --query "{optimization_area}" --limit 3. The{optimization_area}variable is interpolated directly into the command string, creating a vulnerability to shell command injection if the input source is compromised. - [PROMPT_INJECTION]: The skill processes external visual specifications as part of its primary workflow. The ingestion of this untrusted data lacks sanitization or boundary markers, exposing the agent to indirect prompt injection risks.
- Ingestion points: Visual specifications in
SKILL.mdStep 1. - Boundary markers: None present.
- Capability inventory: Access to shell execution (
Bash), file system modification (Write,Edit), and network tools (WebFetch,WebSearch). - Sanitization: No evidence of content sanitization or validation.
Audit Metadata