The Agent Skills Directory

[PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from external websites via the Firecrawl CLI. This creates an attack surface for indirect prompt injection, where malicious instructions embedded in crawled web content could attempt to influence the agent's subsequent actions.
Ingestion points: The firecrawl crawl command in SKILL.md fetches content from arbitrary, user-specified URLs into the agent's workspace.
Boundary markers: The skill does not provide explicit delimiters or instructions to the agent to ignore or isolate instructions found within the retrieved data.
Capability inventory: The skill is authorized to execute firecrawl and npx firecrawl commands via Bash to interact with the file system and network.
Sanitization: No sanitization, filtering, or validation of the retrieved web content is performed by the skill instructions.
[COMMAND_EXECUTION]: The skill uses Bash to execute firecrawl and npx firecrawl commands. These commands are restricted via the allowed-tools configuration to their specific CLI patterns.
[EXTERNAL_DOWNLOADS]: The skill suggests the use of npx firecrawl, which involves downloading and executing the Firecrawl package from the public npm registry at runtime. Firecrawl is recognized as a well-known service for web scraping.

firecrawl-crawl