The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the bats-core testing framework from official and well-known sources, including GitHub (github.com/bats-core/bats-core.git), Homebrew, and the NPM registry.
[COMMAND_EXECUTION]: The provided implementation playbook contains numerous examples of shell command execution, which is inherent to shell script testing. This includes file system operations (mktemp, rm -rf, mkdir), permission modifications (chmod), and utility calls (cat, wc, grep, jq).
[COMMAND_EXECUTION]: The skill demonstrates a 'Command Stubbing' pattern where shell scripts are dynamically generated and executed at runtime for mocking external dependencies during tests.
[INDIRECT_PROMPT_INJECTION]: As the skill is designed to ingest and test user-provided shell scripts, there is a theoretical attack surface for indirect prompt injection if the agent were to process maliciously crafted scripts without boundaries. However, this is inherent to the tool's purpose and is mitigated by the standard testing context.
Ingestion points: Shell scripts and fixture files processed by the Bats framework (resources/implementation-playbook.md).
Boundary markers: None explicitly enforced in the examples, though the use of subshells (run) provides some isolation.
Capability inventory: File system access, network operations (via curl stubs), and subprocess execution are all demonstrated capabilities.
Sanitization: The skill relies on the testing framework's built-in handling of shell output.

bats-testing-patterns