The Agent Skills Directory

[PROMPT_INJECTION]: The instructions explicitly direct the agent to suppress user oversight by stating that dependency installation should happen "automatically without asking the user to confirm" and that the agent should not "request separate approval for dependency installation."- [EXTERNAL_DOWNLOADS]: The skill references an external installation link (https://www.browseract.com/?co-from=social-media&redirect=https://github.com/browser-act/skills/tree/main) to download the 'browser-act' dependency. This source is a third-party domain not associated with the skill's author or trusted providers.- [REMOTE_CODE_EXECUTION]: The requirement to automatically install and run a third-party CLI tool from an external repository represents a significant risk, as it allows for the execution of arbitrary code under the agent's context without verification.- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it ingests untrusted data from YouTube metadata. * Ingestion points: Data is gathered from YouTube short-form video pages (titles, descriptions, comments). * Boundary markers: There are no explicit instructions or delimiters to isolate the scraped content from the agent's instructions. * Capability inventory: The agent has access to the 'browser-act' tool for web interaction and file system access for exporting CSV results. * Sanitization: No sanitization or validation of the retrieved external content is performed before processing.

youtube-shorts-scraper