The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions reference the @different-ai/opencode-browser package for execution via npx. This package is a vendor-owned resource associated with the skill's author.
[COMMAND_EXECUTION]: The documentation includes instructions for running browser automation tools and checking status via command-line interface calls.
[INDIRECT_PROMPT_INJECTION]: The skill interacts with external web pages, creating a potential surface for indirect prompt injection from untrusted web content.
Ingestion points: External web content is processed via browser_snapshot, browser_list, and browser_eval as described in SKILL.md.
Boundary markers: There are no explicit instructions for using delimiters to isolate untrusted web data.
Capability inventory: The skill possesses browser-level capabilities including navigation, element interaction (browser_click, browser_fill), and script evaluation (browser_eval).
Sanitization: The instruction set does not specify procedures for sanitizing or validating the content retrieved from browsers.

browser-automation