The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs the @membranehq/cli package from the official NPM registry. This tool is the primary interface for the Membrane platform and is a trusted dependency provided by the vendor. \n- [COMMAND_EXECUTION]: The skill utilizes the membrane CLI for authentication (membrane login), creating connections (membrane connect), and executing API actions (membrane action run). These are standard operational commands for the platform's intended integration workflow. \n- [CREDENTIALS_UNSAFE]: No hardcoded API keys or secrets were detected. The skill correctly instructs the user to use the vendor's secure OAuth-based authentication flow, ensuring that credentials are managed server-side rather than in local files. \n- [PROMPT_INJECTION]: The skill includes an attack surface for indirect prompt injection by accepting natural language queries in the --intent and action create command parameters. 1. Ingestion points: membrane action list --intent \"QUERY\" and membrane action create \"DESCRIPTION\" in SKILL.md. 2. Boundary markers: Not present; the skill interpolates user-provided text directly into CLI arguments. 3. Capability inventory: Execution of arbitrary vendor-defined actions and network operations via the CLI. 4. Sanitization: The skill instructions do not specify sanitization; it is expected to be handled by the Membrane platform logic. \n- [SAFE]: The YAML frontmatter description incorrectly lists CRM entities (Leads, Deals, Pipelines) for Pulsetic, which is a website monitoring tool. This is identified as a non-malicious documentation artifact from a template.

pulsetic