jr-rails-bootstrap
Pass
Audited by Gen Agent Trust Hub on Jul 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches the Homebrew installation script from the official Homebrew repository on GitHub. It also uses npx to download and install a collection of skills from the author's own repository (julianrubisch/skills).
- [REMOTE_CODE_EXECUTION]: The skill executes the official Homebrew installation script directly via a shell pipe using /bin/bash -c.
- [COMMAND_EXECUTION]: The skill runs multiple shell commands to manage system packages and development tools. It also modifies the user's shell configuration files (~/.zprofile and ~/.zshrc) to configure the environment for Homebrew and mise.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by reading and executing commands defined in a project's CLAUDE.md file to start the development server.
- Ingestion points: Project-specific CLAUDE.md file (Phase G, Step 18).
- Boundary markers: Absent; the content is read and used to determine execution commands without explicit delimiters.
- Capability inventory: Shell command execution (including project-defined server commands), file writes (SETUP_SUMMARY.md), and repository management via gh and glab CLIs.
- Sanitization: Absent; the skill relies on the command string found within the external file.
Audit Metadata