The Agent Skills Directory

[COMMAND_EXECUTION]: The catalog.json file defines an install.command field that executes a shell command (curl -s https://ethskills.com/building-blocks/SKILL.md). This represents execution of shell logic during the installation phase.
[REMOTE_CODE_EXECUTION]: The setup section in catalog.json instructs the agent/user to install external plugins from unverified sources, specifically claude plugin install https://github.com/austintgriffith/ethskills. This facilitates the execution of remote code from a third-party repository.
[EXTERNAL_DOWNLOADS]: The skill points to and downloads data from external domains (ethskills.com and github.com/austintgriffith/*) that are not recognized as trusted or well-known technology providers.
[PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by explicitly instructing the agent to read and follow guidelines from a remote URL (https://ethskills.com/building-blocks/SKILL.md).
Ingestion points: Remote markdown file at ethskills.com.
Boundary markers: None provided in the instructions to isolate the remote content.
Capability inventory: The skill allows shell command execution via curl and installation of external plugins.
Sanitization: No validation or sanitization is performed on the content retrieved from the external URL before the agent processes it.

Building Blocks