The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides various shell commands for developer auditing and test execution.
Evidence: references/fix-strategies.md includes Go, Python, and Node.js test runner commands (e.g., go test -count=100, pytest --count=20, npx jest --repeat=20) and a bash script using grep to identify tests lacking assertions.
[EXTERNAL_DOWNLOADS]: The reference materials suggest the use of established third-party testing utilities.
Evidence: references/fix-strategies.md provides instructions to install pytest-repeat and pytest-randomly via pip. These are well-known community plugins for the pytest framework.
[INDIRECT_PROMPT_INJECTION]: The skill processes external repository content, which presents a surface for indirect prompt injection.
Ingestion points: The skill is designed to read AGENTS.md, production source code, and existing test files within the target repository (SKILL.md).
Boundary markers: Instructions explicitly require a pre-defined 'brief' that identifies owned files and concrete findings before the agent begins edits.
Capability inventory: The skill utilizes test runners and file system search tools (grep) across multiple languages.
Sanitization: Behavior is constrained by the requirement to preserve observable behavior and follow specific patterns documented in the provided catalogs.

testing-preferred-patterns