The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes @playwright/cli to interact with browser environments and executes an inline Python script using the standard csv and os modules to write testing scenarios to a local CSV file in the scenarios/ directory.
[EXTERNAL_DOWNLOADS]: The skill interacts with external websites provided by the user to perform reconnaissance. It explicitly forbids completing transactions, creating accounts, or submitting forms that have real-world side effects during the mapping phase.
[PROMPT_INJECTION]: The skill processes untrusted data from external websites, which presents an indirect prompt injection surface.
Ingestion points: Live website content accessed via Playwright tools (SKILL.md).
Boundary markers: The skill lacks explicit boundary markers for external content but instructs the agent to 'look, don't finalize' to limit impact.
Capability inventory: Browser interaction via Playwright and file-writing capabilities via Python.
Sanitization: No explicit sanitization of discovered website content is mentioned before it is processed into the final CSV output.

scenario-mapper