The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides C and Python source files and instructs the agent to compile and execute them for verification. Specifically, it uses gcc for compiling fixtures such as lock_order_pthread.c and abba_deadlock_pthread.c, and python to run cycle detection models. These operations are performed on the skill's own internal files.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to review kernel source code, which constitutes an ingestion point for potentially untrusted data that could contain malicious instructions.
Ingestion points: The agent reviews kernel code provided in the user context, as specified in the SKILL.md description and use cases.
Boundary markers: The instructions do not specify the use of delimiters or warnings to prevent the agent from following instructions embedded in the code being reviewed (e.g., in code comments).
Capability inventory: The skill leverages powerful tools including gcc, python, make, and qemu-system-x86_64 for building and testing code.
Sanitization: There are no instructions for sanitizing or filtering external code before it is processed or executed.

deadlock-kernel-prevention