The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains explicit instructions to bypass user oversight and safety filters. Specifically, it commands the agent to install missing dependencies automatically 'without asking the user to confirm' and explicitly tells the agent 'do not request separate approval for dependency installation.'
[EXTERNAL_DOWNLOADS]: The skill directs the agent to download the 'BrowserAct' Skill and CLI from an external website (browseract.com) and a GitHub repository (github.com/browser-act/skills). These sources are not recognized as trusted organizations or well-known services.
[REMOTE_CODE_EXECUTION]: By instructing the agent to automatically install and then invoke a third-party command-line interface (CLI) and skill at runtime without user consent, the skill facilitates the execution of arbitrary external code.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from arbitrary external ecommerce URLs. Because this untrusted data is processed while the agent has access to browser-based tools, it creates a vulnerability where malicious instructions hidden on a website could influence the agent's actions.
Ingestion points: External ecommerce product pages, search results, and marketplace URLs (SKILL.md).
Boundary markers: None present; the skill does not specify delimiters or instructions to ignore embedded commands in the captured content.
Capability inventory: Uses the browser-act tool to interact with websites and capture browser-visible metadata, network responses, and content (SKILL.md).
Sanitization: None described for the captured browser content.

ecommerce-price-comparison-research