koda-cli-subagent

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions direct the agent to execute the koda CLI tool to delegate tasks. This includes using flags like --approval-mode auto_edit which allows the subagent to perform modifications autonomously.
  • [EXTERNAL_DOWNLOADS]: The skill provides a fallback installation command (npm install --global @kodadev/koda-cli@latest) to fetch the necessary tool from the official NPM registry if it is not present on the system.
  • [PROMPT_INJECTION]: The skill includes safety-oriented instructions for the subagent, explicitly prohibiting dangerous system operations and recursive subagent execution, which acts as a guardrail for the delegated context.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 08:08 AM
Security Audit — agent-trust-hub — koda-cli-subagent