The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install external binaries using npx (@mvanhorn/printing-press-library) and go install (github.com/mvanhorn/printing-press-library). These resources originate from the identified vendor.
[DATA_EXFILTRATION]: The CLI tool includes a --deliver webhook:<url> feature that allows routing command output to arbitrary remote URLs via HTTP POST. While a functional feature, it provides a direct mechanism for data exfiltration if the agent is directed to send sensitive information.
[COMMAND_EXECUTION]: The skill is designed to drive the pokeapi-pp-cli binary by constructing shell commands from user-supplied arguments. This creates a potential command injection surface if the agent does not properly sanitize inputs before execution.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user data via command-line arguments (e.g., in the which command and SQL queries) and lacks explicit boundary markers or sanitization instructions to prevent embedded malicious instructions from influencing behavior.
Ingestion points: User-supplied strings in $ARGUMENTS and natural language queries passed to the which subcommand.
Boundary markers: Absent. The instructions do not specify the use of delimiters or warnings to ignore instructions embedded in the data.
Capability inventory: Execution of shell commands via the Bash tool and network transmission via the webhook sink.
Sanitization: Absent. There are no instructions for escaping or validating input before interpolation into shell commands.
[DYNAMIC_EXECUTION]: The installation procedures involve go install, which downloads source code from a remote repository and compiles it into an executable binary on the local system at runtime.

pp-pokeapi