industry-trends
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from competitor ad libraries and transcripts, creating a vulnerability where embedded instructions could manipulate agent output.\n
- Ingestion points: External data enters the context via
mcp__motion__get_creative_transcript,mcp__motion__get_inspo_creatives, andmcp__motion__get_inspo_brand_context.\n - Boundary markers: The skill lacks explicit delimiters or instructions to ignore potential commands within the ingested competitor data.\n
- Capability inventory: Execution is limited to read-only vendor tools; there are no capabilities for file modification or arbitrary command execution.\n
- Sanitization: There is no evidence of content filtering or validation for the external transcripts and creative metadata.
Audit Metadata