The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill instructs the user to use sudo to create a directory and apply chmod 777 (world-writable) permissions to a path within the Python site-packages directory. Granting such broad permissions to system-level library directories is a dangerous practice that can be exploited for privilege escalation or persistence by allowing unprivileged processes to write to sensitive locations.
[COMMAND_EXECUTION]: The skill facilitates the execution of various system commands and utilities including chipsec_main, chipsec_util, find, mkdir, and chmod. While these are part of the intended firmware analysis workflow, the inclusion of sudo for directory modification increases the risk profile.
[INDIRECT_PROMPT_INJECTION]: The skill includes workflows that ingest untrusted external data (firmware files extracted via binwalk or found via find) and processes them using shell loops and the chipsec suite. This creates an attack surface where maliciously crafted filenames or firmware metadata could attempt to exploit the processing tools or the shell environment.
Ingestion points: Path inputs <firmware.bin>, and files discovered via find in the filesystem.
Boundary markers: None identified in the provided scripts to delimit untrusted data from shell commands.
Capability inventory: The skill uses subprocess calls to execute firmware analysis tools and filesystem commands.
Sanitization: There is no evidence of filename sanitization or escaping when processing files discovered at runtime, making it vulnerable to command injection through malicious filenames.

chipsec