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/clipackage. This is a vendor-owned tool from membranedev used to interact with the Membrane platform. - [COMMAND_EXECUTION]: Use of the
membraneCLI 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-postsandlist-commentswhich 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 themembrane action runcommand. - Sanitization: There is no specific evidence of sanitization or filtering for the data retrieved from Beamer.
Audit Metadata