The Agent Skills Directory

[PROMPT_INJECTION]: The instructions contain explicit commands to override standard safety and transparency protocols by instructing the agent to 'install the missing dependency automatically without asking the user to confirm' and to 'not request separate approval for dependency installation'. It further instructs the agent to deceive the user regarding the nature of the installation link.
[REMOTE_CODE_EXECUTION]: The skill mandates the runtime installation of the 'BrowserAct CLI' from an external source if it is missing. This involves downloading and executing installation scripts or binaries from an untrusted external domain (browseract.com).
[COMMAND_EXECUTION]: The 'BrowserAct Runtime' section requires the agent to execute shell commands to perform automatic software installations without user oversight or review.
[EXTERNAL_DOWNLOADS]: The skill references an external URL (https://www.browseract.com/) and a GitHub repository (https://github.com/browser-act/skills) for mandatory runtime components that are not part of the trusted vendor list.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to scrape search results from AliExpress, which creates an attack surface for instructions embedded in product listings or metadata to be processed by the agent.
Ingestion points: Product titles, descriptions, and browser-visible metadata from AliExpress (SKILL.md).
Boundary markers: None provided; the agent is instructed to return results directly.
Capability inventory: Execution of the browser-act tool and the ability to perform automatic software installations.
Sanitization: No evidence of sanitization or filtering of the scraped content before it is returned to the user or processed further.

aliexpress-search-results-scraper