beamer

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires the installation of the @membranehq/cli package. This is a vendor-owned tool from membranedev used to interact with the Membrane platform.
  • [COMMAND_EXECUTION]: Use of the membrane CLI for managing connections and executing actions is the intended method for this skill's functionality. Authentication is handled server-side by the vendor.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface (Category 8).
  • Ingestion points: Data enters the agent context through actions like list-posts and list-comments which fetch content from the Beamer API.
  • Boundary markers: No explicit markers or delimiters are defined in the instruction templates.
  • Capability inventory: The skill allows writing data back to Beamer (e.g., create-post, delete-comment) using the membrane action run command.
  • Sanitization: There is no specific evidence of sanitization or filtering for the data retrieved from Beamer.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 06:30 AM