The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Installs the Membrane CLI tool (@membranehq/cli) globally from the npm registry. This package is provided by the author to enable the skill's integration capabilities.
[COMMAND_EXECUTION]: Executes shell commands via the membrane CLI to authenticate, manage JFrog connections, and run automation actions.
[PROMPT_INJECTION]: The skill processes data retrieved from external JFrog instances, presenting a surface for indirect prompt injection.
Ingestion points: Retrieves artifact properties, build metadata, and release bundle information from JFrog via the membrane action run command.
Boundary markers: Instructions do not specify delimiters or isolation markers for data retrieved from external sources.
Capability inventory: The agent can execute shell commands using the membrane binary and install Node.js packages via npm.
Sanitization: No explicit validation or filtering of content fetched from JFrog is described before it is processed by the agent.
Mitigation: Wrap data retrieved from JFrog in distinct delimiters and instruct the agent to ignore any embedded instructions. Implement schema validation for any content returned by external actions.

jfrog