The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill documents the opencli plugin install github:user/repo command, which allows the installation and execution of code from arbitrary GitHub repositories. This enables the execution of unverified third-party logic within the user's environment.
[COMMAND_EXECUTION]: The opencli external install and opencli external register features allow the skill to install and invoke arbitrary command-line tools, inheriting the user's permissions.
[DATA_EXFILTRATION]: The tool supports strategies like COOKIE and INTERCEPT that capture authentication tokens and cookies from active, logged-in Chrome sessions to automate site interactions. While this is the intended purpose, it represents access to highly sensitive session data.
[EXTERNAL_DOWNLOADS]: The skill instructions involve downloading the core utility from the NPM registry (@jackwener/opencli) and cloning source code from the vendor's GitHub repository (github.com/jackwener/OpenCLI).
[PROMPT_INJECTION]: The tool provides an automated path for the agent to ingest data from arbitrary websites, creating a significant attack surface for indirect prompt injection.
Ingestion points: Data retrieved from external websites and Electron apps via opencli <site> <command>.
Boundary markers: None identified in the instruction set to separate untrusted web content from agent instructions.
Capability inventory: The skill can execute shell commands (opencli), write to the home directory (~/.opencli), and install/run third-party plugins.
Sanitization: No explicit sanitization or filtering of website content is documented before the data is processed by the agent.

opencli-usage