The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes content from various file formats (PDF, DOCX, HTML, etc.), which creates an attack surface for indirect prompt injection if the files contain adversarial instructions.\n
Ingestion points: Untrusted data enters the agent's context through the firecrawl parse command which reads local files.\n
Boundary markers: There are no boundary markers or explicit instructions provided to the agent to ignore or delimit instructions found within the parsed file content.\n
Capability inventory: The skill possesses the capability to execute shell commands and write data to the filesystem.\n
Sanitization: No evidence of sanitization, filtering, or validation is provided to mitigate the risk of instructions embedded in the documents being followed by the agent.\n- [EXTERNAL_DOWNLOADS]: The skill configuration allows for the execution of npx firecrawl-cli, which involves downloading the vendor's package from the npm registry.\n
Evidence: The allowed-tools section in the frontmatter explicitly permits Bash(npx firecrawl-cli *).\n- [COMMAND_EXECUTION]: The skill's primary functionality is implemented through shell command execution.\n
Evidence: The allowed-tools include Bash(firecrawl *) and Bash(npx firecrawl-cli *) for document parsing and credit usage monitoring.

firecrawl-parse