design-team

Warn

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill utilizes 'npx' to download and execute the 'designteam' package from the npm registry for project orchestration. It also allows generating shareable links at 'designteam.app'.
  • [COMMAND_EXECUTION]: Project management steps, including planning and task checkouts, are performed using shell commands via the vendor-provided CLI tool.
  • [DYNAMIC_EXECUTION]: Multiple agents (design-engineer.md, qa-lead.md) are granted access to the 'Bash' tool to implement code and run audits. The 'designteam run' command specifically allows for the execution of arbitrary scripts through the '--command' flag.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: Untrusted data enters the agent context via the 'researcher' agent, which uses 'WebSearch' and 'WebFetch' tools (researcher.md).
  • Boundary markers: The instructions do not define delimiters or warnings to ignore embedded instructions within research data.
  • Capability inventory: The skill possesses capabilities for command execution via 'Bash' and file modification via 'Write'.
  • Sanitization: There is no evidence of input validation or sanitization for external content before it is processed by the team.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 14, 2026, 03:29 PM
Security Audit — agent-trust-hub — design-team