The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires executing make and docker compose commands on the host system to initialize and manage the skillbox development container. Additionally, it executes user-defined build and test commands within the container environment.
[EXTERNAL_DOWNLOADS]: The skill's primary function involves downloading and analyzing source code from external, untrusted "upstream" repositories.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external repositories which may contain malicious content.
Ingestion points: Files from upstream repositories specified by the user or discovered via build-vs-clone.
Boundary markers: No explicit delimiters or warnings are used to isolate untrusted code during analysis.
Capability inventory: Host-side execution of make and docker compose; container-side execution of compilers, test suites, and shell commands.
Sanitization: No explicit sanitization or validation of the untrusted code is performed prior to analysis or execution of tests.

claude-clone