The Agent Skills Directory

[COMMAND_EXECUTION]: The skill involves the execution of numerous security-focused command-line tools such as binwalk, flashrom, qemu, picocom, and apktool. These are used for tasks ranging from hardware interaction (UART/JTAG) to binary simulation and analysis. The instructions include privileged operations such as using sudo for tool installation and chroot for mounting simulated filesystems, which are standard for the described security research activities.
[EXTERNAL_DOWNLOADS]: The documentation references external sources for obtaining firmware images and security research tools. These include vendor support portals and well-known security community repositories on GitHub (e.g., OWASP IoTGoat, DVRF). All referenced sources are typical within the context of firmware security assessment.
[PROMPT_INJECTION]: As the skill is designed to process and analyze untrusted external data (firmware images and their extracted filesystems), it inherently possesses an indirect prompt injection attack surface. This means that if an ingested firmware file contains specially crafted strings (e.g., in filenames or configuration metadata), it could potentially influence the agent's behavior during the analysis process. This is a common risk for tools designed for security auditing of untrusted inputs.

firmware-analysis