The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the membrane CLI to perform operations such as logging in, connecting to search properties, and running actions. This is the intended behavior for interacting with the Membrane platform.
[EXTERNAL_DOWNLOADS]: Instructs the installation of the @membranehq/cli package from the npm registry. This is the official tool provided by the vendor (membranedev) for the platform's functionality.
[REMOTE_CODE_EXECUTION]: Features a capability to create new actions dynamically (membrane action create), which generates integration code on the Membrane platform based on user descriptions.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it retrieves and processes data from external sources (Google Search Console).
Ingestion points: Data such as site URLs, performance reports, and sitemap information enter the agent context via the membrane action run command in SKILL.md.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to distinguish between platform data and potentially malicious strings embedded in site metadata.
Capability inventory: The agent has the ability to execute shell commands and modify resources via the membrane CLI as documented in SKILL.md.
Sanitization: The skill does not specify any sanitization or validation steps for the data retrieved from the Search Console API.

google-search-console