design-import

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted content from design handoff bundles to drive component scaffolding.
  • Ingestion points: Handoff bundles fetched from user-provided URLs or local file paths, including Markdown chat logs (chats/*.md) and HTML prototypes (project/*.html) as described in Phase 1 of SKILL.md.
  • Boundary markers: The instructions lack explicit delimitation or warnings for the agent to ignore potentially malicious embedded instructions within the bundle's chat files or code prototypes.
  • Capability inventory: The skill has access to powerful tools including Bash, Write, Edit, Agent, and WebFetch across its subagent tasks.
  • Sanitization: No evidence of sanitization, validation, or escaping of the ingested text content is present before the data is processed by the claude-design-orchestrator or frontend-ui-developer agents.
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to perform file system operations on the imported bundle content.
  • Evidence: The skill instructs the agent to execute tar -xzf <bin> -C /tmp/<scratch>/ to extract the gzipped tarball during Phase 1.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 02:15 AM
Security Audit — agent-trust-hub — design-import