The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the firecrawl and npx firecrawl command-line tools to perform browser interactions. This involves executing shell commands to automate web tasks based on natural language prompts or provided code snippets.
[INDIRECT_PROMPT_INJECTION]: The skill processes input via the --prompt and --code arguments, which may contain content derived from scraped webpages. This creates a surface for indirect prompt injection where malicious content on a site could attempt to influence the agent's browser actions.
Ingestion points: Data passed to the --prompt and --code flags in the firecrawl interact command, potentially sourced from prior scrapes.
Boundary markers: None identified; the skill relies on the underlying tool's handling of the input.
Capability inventory: Execution of browser actions (click, fill, navigate) and arbitrary script execution (Bash, Python, Node.js) within the browser session via the firecrawl tool.
Sanitization: No specific sanitization or validation of the input text or code is defined within the skill instructions.

firecrawl-interact