The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions to install the firecrawl-cli package via npm and references documentation from the Firecrawl service. These are standard operations for enabling the skill's core functionality.
[INDIRECT_PROMPT_INJECTION]: The skill processes external web content, which serves as a potential vector for indirect prompt injection attacks.
Ingestion points: Web page content and search results retrieved through scrape, search, and crawl commands (SKILL.md).
Boundary markers: The skill documentation enforces the use of the -o flag to direct output to the .firecrawl/ directory, isolating untrusted data from the immediate context window (rules/security.md).
Capability inventory: The skill is configured to use the Bash tool to execute firecrawl and npx firecrawl commands (SKILL.md).
Sanitization: Guidelines specify the use of shell quoting for URLs and suggest using incremental data processing tools like grep, head, and jq to inspect content safely (rules/security.md).

firecrawl