Battlecard
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a potential vulnerability to indirect prompt injection.
- Ingestion points: The skill ingests untrusted data from web search results (Layer 2) and historical meeting transcripts via RAG (Layer 3) as described in
SKILL.md. - Boundary markers: Absent. There are no instructions to use specific delimiters (e.g., XML tags) or protective 'ignore instructions' prefixes when processing external content.
- Capability inventory: The skill utilizes
execute_actionfor CRM data access (get_deal,get_deal_contacts,get_deal_activities) and performs web searches. - Sanitization: Absent. The instructions do not mention filtering, validation, or sanitization of ingested data before synthesis.
- [NO_CODE]: The skill is composed entirely of markdown-based instructions and reference guides; it does not contain any Python, JavaScript, or other executable script files.
Audit Metadata