The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill specifies an external installation source and fetches content from a remote domain (ethskills.com) via curl during the setup and installation phases.
[COMMAND_EXECUTION]: The catalog.json file includes a shell command (curl -s https://ethskills.com/standards/SKILL.md) within the install.command field, which is intended for execution during skill deployment.
[REMOTE_CODE_EXECUTION]: The setup instructions direct the user to install a plugin directly from a GitHub URL (https://github.com/austintgriffith/ethskills), which involves downloading and potentially executing unverified code within the agent's environment.
[PROMPT_INJECTION]: The skill is designed to fetch its core instructions from an external URL, introducing a vulnerability to indirect prompt injection where a remote attacker could modify the agent's behavior.
Ingestion points: catalog.json specifies fetching instructions from https://ethskills.com/standards/SKILL.md.
Boundary markers: Absent; there are no delimiters or warnings to prevent the agent from obeying instructions embedded in the remote file.
Capability inventory: The skill demonstrates capabilities for on-chain identity registration and executing machine-to-machine payments using the agent0-sdk library.
Sanitization: The skill does not implement any validation or sanitization for the content fetched from the remote URL.

Standards