The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill interacts with the api.browser-use.com domain to manage browser sessions and execute tasks. This is a well-known service for browser automation and the interaction is necessary for the skill's stated purpose.
[COMMAND_EXECUTION]: Documentation includes curl commands for API interaction and a gateway configuration command to link the agent to the remote browser. These are standard integration steps.
[CREDENTIALS_UNSAFE]: API keys are managed through configuration variables (skills.entries.browser-use.apiKey) and are not hardcoded within the skill source.
[PROMPT_INJECTION]: The skill processes external data which represents a surface for indirect prompt injection. Ingestion points: User-provided task prompt and web content loaded via the provisioned browser. Boundary markers: None identified in the documentation. Capability inventory: Remote browser control and configuration updates via the gateway tool. Sanitization: Not specified in the skill code. This risk is inherent to browser automation functionality and is noted for situational awareness.

browser-use