The Agent Skills Directory

[SAFE]: The skill instructions utilize the @membranehq/cli tool, which is the official CLI for the skill's author (membranedev). The use of this tool for managing connections and performing API requests is consistent with the skill's stated purpose.
[SAFE]: Authentication is handled through a secure browser-based flow (membrane login), ensuring that sensitive credentials like OAuth tokens are managed by the Membrane platform and not stored insecurely within the skill or environment variables.
[DATA_EXPOSURE_&_EXFILTRATION]: No sensitive local file access or exfiltration patterns were detected. All network operations are directed through the Membrane platform for interacting with the official Google YouTube Analytics API.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it retrieves data from YouTube Analytics (comments, reports). However, this is a standard functional requirement for such an integration.
Ingestion points: Data is ingested through membrane action run and membrane request commands (referenced in SKILL.md).
Boundary markers: No specific boundary markers or 'ignore' instructions are present in the prompt templates.
Capability inventory: The skill uses shell command execution via the membrane CLI to perform API actions.
Sanitization: Sanitization is handled at the platform level by Membrane, which provides structured output schemas for actions.

youtube-analytics