product-design-audit
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands, including a vendor-provided bootstrap script using
nodeandnpm installfor project setup as specified inreferences/local-prototype-preflight.md.- [EXTERNAL_DOWNLOADS]: Performs external package downloads from the NPM registry when initializing local prototypes vianpm install.- [PROMPT_INJECTION]: Contains a vulnerability surface for indirect prompt injection as it processes untrusted data from web pages during audits. - Ingestion points: Web content (DOM and screenshots) captured via OpenCowork Browser and Chrome skills in
SKILL.md. - Boundary markers: No explicit delimiters or instructions are used to isolate untrusted web content from the agent's core instructions.
- Capability inventory: The agent possesses capabilities for file system modification, command execution (node, npm), and browser-based network interactions.
- Sanitization: There is no evidence of sanitization or filtering of external data before it is incorporated into the agent's reasoning context.
Audit Metadata