reverse-engineer

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses various platform-specific CLI tools (such as git, rg, powershell, mdfind, and dpkg-query) to perform discovery of installed applications and to inspect on-disk artifacts. These commands are limited to read-only operations to ensure the integrity of the host system.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the cloning of remote repositories using git clone and gathering documentation via a research agent. These downloads are isolated to user-confirmed scratch directories to prevent contamination of the primary workspace.
  • [SAFE]: The instructions include defensive measures against indirect prompt injection by explicitly directing the agent to treat source files, issues, and web content as untrusted data rather than instructions. Additionally, it explicitly forbids the installation or execution of untrusted repository code.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 10:16 PM
Security Audit — agent-trust-hub — reverse-engineer