The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions direct the agent to install and run code from external sources using npx -y @mvanhorn/printing-press-library and go install github.com/mvanhorn/printing-press-library/.... These resources belong to the vendor's infrastructure.
[COMMAND_EXECUTION]: The skill is built around executing the stac-pp-cli shell command. It includes an --agent flag that automatically bypasses confirmation prompts by including --yes and --no-input.
[DATA_EXFILTRATION]: The CLI provides a --deliver webhook:<url> feature that allows the output of any command to be sent via an HTTP POST request to an arbitrary URL, which could be exploited to exfiltrate information.
[INDIRECT_PROMPT_INJECTION]: The skill processes imagery metadata from external STAC APIs (e.g., AWS Earth Search), which are untrusted data sources.
Ingestion points: Data retrieved from STAC APIs via the items and collections commands.
Boundary markers: None; the instructions do not specify the use of delimiters or instructions to ignore embedded content in the API responses.
Capability inventory: Shell execution of the CLI tool, atomic file writing via --deliver file:<path>, and arbitrary network requests via the webhook sink.
Sanitization: No sanitization or validation of the external API content is described.

pp-stac