The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the webcmd CLI via the Bash(webcmd:*) tool to perform browser automation tasks. This involves executing shell commands for every browser interaction, such as browser state, browser click, and browser type. These operations are consistent with the skill's stated purpose of driving a live Chrome window.
[INDIRECT_PROMPT_INJECTION]: The skill provides an interface for reading and processing external web content, which serves as a potential surface for indirect prompt injection.
Ingestion points: Data is ingested from external sources via browser state, browser extract, browser network, and web read subcommands.
Boundary markers: The CLI returns structured JSON envelopes for command outputs, which helps the agent distinguish between tool metadata and the ingested web content.
Capability inventory: The skill possesses capabilities to interact with the DOM (clicking, typing, form filling), manage browser tabs, and execute local shell commands via the webcmd tool.
Sanitization: There is no evidence of explicit sanitization or filtering of the content retrieved from web pages before it is presented to the agent.
[DYNAMIC_EXECUTION]: The browser eval command allows for the runtime execution of JavaScript within the browser context. The skill instructions specify that this should be used for read-only data extraction and recommend wrapping the code in an IIFE to return JSON. This pattern is a standard feature for advanced browser interaction tools.

webcmd-browser