The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands including curl for target verification and pkill to terminate background tunnel processes during cleanup.
[COMMAND_EXECUTION]: It leverages browser-harness to execute Python-based browser automation scripts via shell heredocs.
[EXTERNAL_DOWNLOADS]: The skill instructs users to install the browser-harness CLI and agent plugins, which are recognized as vendor-specific resources for the author 'browser-use'.
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection as it navigates and processes content from external websites that could contain malicious instructions.
Ingestion points: Target URLs provided by users are accessed via the browser-harness tool in SKILL.md and references/methodology.md.
Boundary markers: The skill lacks explicit instructions or delimiters to isolate the agent from potentially malicious commands embedded in the external website's HTML or console output.
Capability inventory: The agent has the ability to execute shell commands, manage network tunnels, and control cloud browser sessions.
Sanitization: No filtering or sanitization of the retrieved website content is performed before the agent evaluates the page state.

qa