readelf
Warn
Audited by Gen Agent Trust Hub on Jul 8, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The installation scripts use shell commands to manage system packages. Specifically, 'scripts/install.ps1' invokes 'sudo apt-get' to install software on Linux systems, which requires elevated privileges.
- [EXTERNAL_DOWNLOADS]: The skill automates the download and installation of the 'readelf' utility from official system repositories including Homebrew (macOS), APT (Debian/Ubuntu), DNF (Fedora/RHEL), Pacman (Arch), and APK (Alpine). These are well-known, trusted package registries.
- [PROMPT_INJECTION]: The skill is designed to process external, untrusted binary files (ELF executables and libraries). This presents an indirect prompt injection surface where malformed binary metadata could attempt to influence the agent's behavior or exploit vulnerabilities in the 'readelf' parser.
- Ingestion points: The 'readelf' command processes user-provided file paths as documented in 'SKILL.md' and 'references/cheatsheet.md'.
- Boundary markers: The 'SKILL.md' file includes safety notes advising the user to validate files from untrusted sources and to pipe output safely.
- Capability inventory: The skill uses subprocess calls to execute the 'readelf' binary and various package managers ('apt', 'brew', 'dnf', etc.) across 'scripts/install.sh' and 'scripts/install.ps1'.
- Sanitization: No automated sanitization or sandboxing of the input files is implemented; the skill relies on manual validation warnings.
Audit Metadata