vector-embed
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to take untrusted input text or external file content and pass it directly into a shell command line argument via the Bash tool. This creates an indirect prompt injection surface.
- Ingestion points: File content retrieved via the Read tool or text supplied via the argument hint in SKILL.md.
- Boundary markers: Absent. There are no delimiters or instructions to treat embedded content as passive data.
- Capability inventory: Requires the Bash tool, Read tool, and memory store tools to function.
- Sanitization: Absent. The instructions do not direct the agent to escape quotes, dollar signs, or backticks before interpolating the text into the command string.
- [COMMAND_EXECUTION]: The skill explicitly directs the agent to invoke shell commands using the Bash tool to install npm dependencies (
npm install ruvector@0.2.25) and execute CLI binaries vianpxwith dynamically supplied text parameters.
Audit Metadata