The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a framework for ingesting and processing external markdown content from URLs via Firecrawl. This content is intended for agent consumption, creating a vulnerability where a malicious website could include hidden instructions to influence the agent.
Ingestion points: The firecrawl.crawlUrl and firecrawl.scrapeUrl methods in SKILL.md bring external data into the execution flow.
Capability inventory: The skill allows for file writing and network operations across its scripts.
Boundary markers: Absent. The instructions do not specify how the agent should distinguish between its own logic and instructions found within the scraped data.
Sanitization: Absent. The provided validateScrapedContent function filters for content length and error pages but does not perform any sanitization to strip potential prompt injections or executable directives.
[EXTERNAL_DOWNLOADS]: The skill depends on the @mendable/firecrawl-js Node.js package to interface with the scraping service.

firecrawl-policy-guardrails