The Agent Skills Directory

[COMMAND_EXECUTION]: The skill constructs a shell command string using multiple parameters provided by the user or parsed from the environment.
Evidence: SKILL.md describes the behavior of "assembling the ~/.bits-ut/utd remote_test ... command" from inputs such as working_directory, package_path, and files.
Risk: Without explicit sanitization of these string parameters, an attacker could inject shell operators (e.g., ;, &&) to execute arbitrary commands.
[PROMPT_INJECTION]: The skill processes untrusted data from test logs which could contain malicious instructions designed to manipulate the agent's next steps.
Ingestion points: The skill parses JSON streams from the bits-ut tool and extracts the Output field as described in SKILL.md.
Boundary markers: No delimiters or boundary markers are used to separate the external test output from the agent's instructions.
Capability inventory: The skill executes shell commands and is used within the /speckit.verify workflow, which has the capability to modify or implement code.
Sanitization: There is no evidence of sanitization or escaping of the extracted log content before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill relies on a remote testing mechanism that transmits data to external servers.
Evidence: The skill is titled unit-test-remote and utilizes a remote_test command, implying network communication and remote execution of code.

unit-test-remote