The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external content from URLs and user-provided HTML strings, which creates a surface for indirect prompt injection if the rendered content contains instructions designed to influence the agent's behavior.
Ingestion points: The playwright().screenshot(), playwright().pdf(), playwright().content(), and playwright().render() functions in SKILL.md ingest external data.
Capability inventory: The skill has the capability to write to the filesystem and perform network operations through browser rendering.
Boundary markers: The documentation does not specify the use of delimiters or boundary markers for the rendered content.
Sanitization: There is no mention of sanitization or filtering of the external content before it is processed by the renderer.
[COMMAND_EXECUTION]: The skill allows the agent to write rendered files directly to the local filesystem by specifying a destination path.
Evidence: Usage examples such as playwright().screenshot( "https://boxlang.io", "home.png", ... ) and path : "invoice.pdf" demonstrate direct filesystem write capabilities.

bx-playwright-rendering