doyourmagic
Fail
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill instructs the agent to download and execute shell scripts from the vendor's GitHub repository using shell piping (e.g.,
curl | bash). This is used for the initial setup and for promoting generated skills.\n- [PERSISTENCE]: The installation process modifies the agent's global configuration (~/.claude/settings.json) and installs persistent tools and skills in the user's home directory. This allows the skill's logic to remain active across different sessions.\n- [INDIRECT_PROMPT_INJECTION]: The skill analyzes external repositories, creating a surface for indirect prompt injection. \n - Ingestion points: External repositories cloned for analysis (SKILL.md).\n
- Boundary markers: Uses isolated home directories and sandboxing (SKILL.md).\n
- Capability inventory: Includes repository cloning, file reading, and shell command execution (SKILL.md).\n
- Sanitization: Relies on verifying commands against help menus and parsers rather than explicit input sanitization (SKILL.md).\n- [COMMAND_EXECUTION]: The skill performs extensive shell operations to explore and verify external tools, including running installers and checking command outputs.\n- [EXTERNAL_DOWNLOADS]: The skill downloads setup scripts and framework components from the author's GitHub repositories (
rheinmir/setup,rheinmir/dym).
Recommendations
- HIGH: Downloads and executes remote code from: https://raw.githubusercontent.com/Rheinmir/setup/orca/fdk/tools/fdk-kit.sh - DO NOT USE without thorough review
Audit Metadata