battlecard-generator
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources, creating a surface for indirect prompt injection attacks.
- Ingestion points: Phase 1 of the research process utilizes
web_searchandfetch_webpageto ingest content from competitor websites, third-party review sites (G2, Capterra), and social media platforms (Reddit, Twitter). - Boundary markers: There are no explicit instructions to the agent to treat ingested web content as data only or to ignore instructions embedded within that content.
- Capability inventory: The skill has the capability to perform web searches, fetch webpage content, and write files to the local filesystem (saving battlecards to
clients/<client-name>/...). - Sanitization: The instructions do not define any sanitization, filtering, or escaping mechanisms for the data retrieved from external URLs before it is processed by the LLM.
Audit Metadata