skills/zpankz/mcp-skillset/osgrep/Gen Agent Trust Hub

osgrep

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • Prompt Injection (SAFE): No instructions to override agent behavior or bypass safety filters were detected. The content consists of legitimate technical documentation and search patterns.
  • Command Execution (SAFE): The search-validator.sh script executes local commands like osgrep, jq, awk, and bc for the purpose of validating search results. Command arguments are properly quoted to prevent shell injection.
  • Data Exposure & Exfiltration (SAFE): No evidence of credential exposure or unauthorized data transmission was found. The tool operates on local project indexes.
  • Unverifiable Dependencies & Remote Code Execution (SAFE): The README suggests installing osgrep via standard package managers (npm, Homebrew). It does not use unsafe execution methods like piping remote URLs to a shell.
  • Indirect Prompt Injection (SAFE): The skill provides a surface for processing codebase content. However, it uses structured output (JSON) and parsing utilities (jq) which significantly reduce the risk of an agent incorrectly executing instructions found within the code.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:31 PM