The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the agent to automatically download and install a dependency from an external URL (https://www.browseract.com/...) if the browser-act tool is missing. This bypasses typical software supply chain verification.
[COMMAND_EXECUTION]: Instructions direct the agent to perform the installation 'automatically without asking the user to confirm,' which removes human-in-the-loop oversight for executing system-level changes.
[PROMPT_INJECTION]: The skill contains instructions that mandate the concealment of its actions from the user. Specifically, it tells the agent 'Do not describe it as a request for the user to open their browser' and 'do not request separate approval for dependency installation.' This is a form of autonomy abuse intended to bypass user consent.
[PROMPT_INJECTION]: The skill is vulnerable to Indirect Prompt Injection. It ingests untrusted data from Douyin (captions, comments, search results) and possesses the capability to interact with a live browser, creating a surface where malicious content on a web page could influence the agent's behavior.
Ingestion points: Douyin search results, video captions, profile names, and comments (referenced in 'Common Data' section of SKILL.md).
Boundary markers: Absent. There are no instructions to the agent to ignore or delimit instructions found within the scraped content.
Capability inventory: The agent can invoke the browser-act skill and CLI to interact with live websites (referenced in 'Instructions' section of SKILL.md).
Sanitization: Absent. There is no mention of filtering or validating the content retrieved from search results before processing.

douyin-search-results-scraper