amazon-keyword-research
Fail
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: HIGHDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DYNAMIC_EXECUTION]: The bash script
scripts/research.shis susceptible to Python code injection. It constructs a Python command string by directly interpolating the shell variable$KEYWORDinto single quotes within apython3 -ccall. A maliciously crafted keyword containing a single quote can escape the string literal and execute arbitrary Python code in the skill's environment. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from external search results and user keywords without sufficient sanitization.
- Ingestion points: User keywords and competitive data fetched from Amazon and Google Trends as described in
SKILL.md. - Boundary markers: No explicit delimiters or instructions are provided to the agent to distinguish external data from core instructions.
- Capability inventory: Local script execution and network search capabilities.
- Sanitization: The skill lacks logic to escape or validate content retrieved from external sources before presenting it to the agent or user.
- [EXTERNAL_DOWNLOADS]: The skill fetches autocomplete suggestions and market trends from Amazon's official APIs and Google Trends. These are well-known and appropriate services for the skill's keyword research purpose.
Recommendations
- AI detected serious security threats
Audit Metadata