The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documents the opencli plugin install github:user/repo command, which enables the installation of third-party code from arbitrary, unverified GitHub repositories.
[COMMAND_EXECUTION]: The skill instructs the agent to use the Bash tool to run opencli commands, which includes functionality to install external tools via system package managers like brew, apt, and npm through the opencli external install interface.
[DYNAMIC_EXECUTION]: The skill supports the creation and execution of custom JavaScript adapters and the registration of arbitrary external binaries (e.g., via opencli external register), which are executed at runtime.
[INDIRECT_PROMPT_INJECTION]: The skill serves as a surface for indirect prompt injection as it facilitates the ingestion of data from untrusted external websites and applications.
Ingestion points: Untrusted data enters the agent context through opencli <site> <command> and opencli browser extraction commands described in SKILL.md.
Boundary markers: The instructions do not define explicit boundary markers or delimiters for the data returned from external sites.
Capability inventory: The skill possesses capabilities for shell command execution (Bash), file reading (Read), and automated browser interaction.
Sanitization: No specific sanitization or filtering logic is documented for the content retrieved from external sources before it is processed by the agent.

opencli-usage