grok-search

Fail

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill instructions provide a command curl -fsSL https://github.com/patrick-fu/grok-search-cli/releases/latest/download/install.sh | bash to install a local dependency. Executing remote scripts directly in a shell is a dangerous practice that can lead to system compromise. While the agent is instructed not to run it automatically, presenting this as the installation method poses a significant risk from a non-verified source.
  • [COMMAND_EXECUTION]: The skill executes the grok-search command in a local shell. It passes user queries using a shell pipe, which could be vulnerable to command injection if the agent interpolates unsanitized user input into the shell command string.
  • [INDIRECT_PROMPT_INJECTION]: 1. Ingestion points: The skill fetches search results from public web pages and X posts. 2. Boundary markers: The instructions lack markers or warnings to the agent to disregard instructions found within the retrieved content. 3. Capability inventory: The skill has access to local shell execution and can trigger updates or suggested installations. 4. Sanitization: No sanitization process is described for the content returned by the search tool.
Recommendations
  • HIGH: Downloads and executes remote code from: https://github.com/patrick-fu/grok-search-cli/releases/latest/download/install.sh - DO NOT USE without thorough review
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Aug 29, 2026, 09:51 PM
Security Audit — agent-trust-hub — grok-search