AI Ark Semantic Search

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: No malicious instructions, prompt injections, or security bypasses were found in the skill configuration.- [NO_CODE]: The skill consists solely of a markdown definition (SKILL.md) and does not include any executable scripts or binary files, significantly reducing the attack surface.- [DATA_EXFILTRATION]: No evidence of unauthorized data transmission was found. The skill interacts exclusively with the expected AI Ark API for company enrichment purposes.- [PROMPT_INJECTION]: The skill processes user-provided natural language queries, creating a potential surface for indirect injection.
  • Ingestion points: natural_language_query input in SKILL.md.
  • Boundary markers: None provided in the prompt logic.
  • Capability inventory: Limited to ai_ark_api calls.
  • Sanitization: Not explicitly defined in the skill markdown.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 11:37 AM