NYC

presentation-architect

Pass

Audited by Gen Agent Trust Hub on Feb 15, 2026

Risk Level: LOWPROMPT_INJECTIONNO_CODE
Full Analysis
  • Indirect Prompt Injection (LOW): The skill processes untrusted user-provided briefs to structure narratives and design slides, creating a potential surface for indirect prompt injection.
  • Ingestion points: User-provided topics, purposes, and constraints gathered in Workflow Step 1.
  • Boundary markers: Absent; the skill lacks instructions to delimit user input from agent instructions.
  • Capability inventory: Limited to Markdown text generation and saving output to .md files; no network or shell execution capabilities identified.
  • Sanitization: Absent; the agent is directed to use user content directly to fulfill design specifications.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 15, 2026, 10:08 PM