The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the @membranehq/cli package from the npm registry. This is the official command-line interface for the Membrane platform and is expected for the skill's functionality.
[COMMAND_EXECUTION]: The skill provides instructions for running multiple CLI commands using the membrane utility, including membrane login, membrane search, membrane connect, and membrane action run. These commands are used to manage scraping workflows and execute actions.
[PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it ingests untrusted data from external websites via web scraping.
Evidence Chain for Indirect Injection:
Ingestion points: Web page content extracted by Firecrawl (SKILL.md).
Boundary markers: Absent; the skill does not specify delimiters or instructions for the agent to ignore potentially malicious content within scraped data.
Capability inventory: The agent can execute shell commands through the Membrane CLI (membrane action run, membrane request).
Sanitization: No explicit sanitization or validation of the scraped data is described in the provided instructions.

firecrawl