The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The README documentation provides a one-line command to install the skill by downloading a shell script from a remote URL (https://raw.githubusercontent.com/nickperkins/testing-skill/main/install.sh) and piping it directly into the bash shell. This executes unverified remote code on the host machine.
[COMMAND_EXECUTION]: The install.sh script executes shell commands to identify active agent skill directories and copy the skill contents. It creates directories and removes existing versions of the skill folder within user-specific paths like ~/.claude/skills and ~/.copilot/skills.
[EXTERNAL_DOWNLOADS]: The skill's setup process fetches the installation script from an external GitHub repository belonging to the author.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it instructs the agent to ingest and adapt to existing code, test files, and configuration data from the user's project.
Ingestion points: Reads source code files, existing test suites, and project configuration files (e.g., package.json, pytest.ini, phpunit.xml) to mirror existing coding patterns.
Boundary markers: The instructions do not provide delimiters or specific guidelines to ignore potentially malicious instructions embedded within the code or configuration files being analyzed.
Capability inventory: The skill is used by agents that typically have file system access and the ability to execute terminal commands to run test runners (e.g., Jest, Pytest).
Sanitization: There are no documented steps for validating or sanitizing the content of external project files before they are processed by the agent.

testing