competitive-battlecard
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection because it is designed to ingest, process, and summarize content from external sources.
- Ingestion points: Content is retrieved from the file system (folders such as
competitive/,sales/,marketing/), user-provided references (via@filesyntax), and tool outputs from internal commands (/context-recall,/feedback-recall). - Boundary markers: The instructions do not define specific delimiters or "ignore instructions" warnings for the model when it processes these external inputs.
- Capability inventory: The skill has the capability to read files, write new markdown files to the project directory, and invoke platform-specific tools for context management and SEO analysis.
- Sanitization: No explicit validation, escaping, or filtering of the ingested content is specified in the methodology or instructions.
Audit Metadata