The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute system commands including apt-get, brew, and docker. It specifically includes commands using sudo to acquire administrative privileges for package installation, which can compromise host system integrity.
[EXTERNAL_DOWNLOADS]: The skill includes instructions to fetch the hadolint binary from its official GitHub repository if system packages are unavailable.
[REMOTE_CODE_EXECUTION]: The installation workflow involves downloading an external binary, placing it in a system-wide executable path (/usr/local/bin/), and granting it execution permissions (chmod +x), which allows for the execution of code from a remote source.
[PROMPT_INJECTION]: The skill processes Dockerfile content from the project repository, creating an indirect prompt injection surface where malicious instructions embedded in a Dockerfile could influence agent behavior.
Ingestion points: Processes Dockerfiles and .hadolint.yaml configuration files found in the project.
Boundary markers: No explicit delimiters or isolation instructions are provided for the untrusted input content.
Capability inventory: The skill utilizes shell commands, system package managers, and container runtimes.
Sanitization: No sanitization or validation of the input file content is performed prior to processing by the tool.

hadolint