The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from external websites into the agent's context, which could potentially contain malicious instructions aimed at influencing the agent's behavior.
Ingestion points: Web search results, page excerpts, and full-page content retrieved via firecrawl search --scrape and firecrawl scrape (SKILL.md).
Boundary markers: The skill does not provide specific instructions for the agent to use delimiters or ignore instructions within the retrieved content.
Capability inventory: The agent has access to shell execution via the Bash tool, allowing it to run CLI tools and file system operations (SKILL.md).
Sanitization: No explicit sanitization or filtering of the external web content is defined within the skill.
[COMMAND_EXECUTION]: The skill instructions direct the agent to execute shell commands to interact with the Firecrawl API and process data.
Evidence: Commands include firecrawl search, firecrawl list, firecrawl scrape, and the use of jq for processing JSON results stored in the .firecrawl/ directory.
[EXTERNAL_DOWNLOADS]: The skill utilizes the npx utility to download and run the vendor's CLI package.
Evidence: The allowed-tools and instructions reference npx firecrawl-cli, which downloads the package from the NPM registry at runtime.

firecrawl-search