The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/makefile_builder.py uses the subprocess module to execute make, gmake, or mingw32-make and compiler toolchains. This behavior is standard and necessary for a build automation skill.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and parses Makefile files from the project workspace. Because make is a powerful tool that can execute arbitrary shell commands defined within those files, there is a theoretical surface for indirect prompt injection if a workspace contains a malicious Makefile. However, this risk is inherent to all build systems and the skill provides no unusual escalation paths.
Ingestion points: Reads Makefile, makefile, or GNUmakefile from the user's workspace in scripts/makefile_builder.py.
Boundary markers: None implemented for the content of the Makefile itself.
Capability inventory: Uses subprocess.run to execute make and toolchain commands in scripts/makefile_builder.py.
Sanitization: Makefile variables are parsed via regex, but the actual build execution relies on the integrity of the project's own Makefile.

build-makefile