The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted content from external URLs, which creates a surface for indirect prompt injection attacks where a malicious website could attempt to influence the agent's behavior.
Ingestion points: External websites monitored via the --page, --scrape-urls, and --crawl-url flags in SKILL.md.
Boundary markers: The instructions do not define specific delimiters or instructions for the agent to ignore potentially malicious commands embedded in the monitored web content.
Capability inventory: The skill can perform network exfiltration via webhooks (--webhook-url), send emails (--email), and execute local CLI commands.
Sanitization: The skill relies on an 'AI judge' for relevance and noise filtering, but it does not implement security-specific sanitization or validation of the monitored data to prevent injection.
[COMMAND_EXECUTION]: The skill uses the firecrawl CLI and npx firecrawl commands to interact with the monitor service. These are allowed vendor-specific tools used for their intended purpose.
[EXTERNAL_DOWNLOADS]: The skill utilizes npx firecrawl, which dynamically fetches the Firecrawl package from the npm registry. As this is a well-known service and the package belongs to the official vendor, this is documented as standard functionality.

firecrawl-monitor