The Agent Skills Directory

[COMMAND_EXECUTION]: The CoverageAnalyzer class uses subprocess.run to execute the test_command string. This creates a potential command injection vulnerability if the command string is derived from untrusted input without proper sanitization.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes local source code, extracting metadata such as docstrings and function decorators to inform test generation. It lacks boundary markers to distinguish between code to be tested and potential instructions embedded within comments or docstrings.
Ingestion points: Local filesystem access via TestGenerator._analyze_python which reads files using open(file_path).
Boundary markers: None identified; the skill processes raw file content and extracted docstrings directly.
Capability inventory: Capability to read local files and execute shell commands (subprocess.run).
Sanitization: No validation or sanitization is performed on the data extracted from source code files before it is processed by the agent.

unit-testing-test-generate