ui-cloner-synthesis

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by ingesting untrusted data without sanitization.\n
  • Ingestion points: Reads project data from plans/01-site-dna.md and user-provided brand interview answers.\n
  • Boundary markers: Absent. Rule 2 and Rule 2b explicitly instruct the agent to transplant artifacts from the source files verbatim or directly.\n
  • Capability inventory: The skill writes the generated output to a local file (plans/03-replication-prompt.md).\n
  • Sanitization: No escaping or validation is performed on the ingested data before it is incorporated into the final output prompt.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 05:19 AM