The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the official Membrane CLI from npm.
Evidence: npm install -g @membranehq/cli@latest in SKILL.md.
Source: Official npm registry, belonging to the skill's vendor (Membrane).
[COMMAND_EXECUTION]: The skill uses the membrane command-line tool to perform operations.
Evidence: membrane login, membrane connect, membrane action list, and membrane action run in SKILL.md.
Purpose: These commands are used for legitimate authentication and to interact with the Docmosis API through the Membrane gateway.
[PROMPT_INJECTION]: The skill provides an attack surface for indirect prompt injection via user-controlled intent and description strings.
Ingestion points: membrane action list --intent "QUERY" and membrane action create "DESCRIPTION" in SKILL.md.
Boundary markers: Absent. The instructions do not specify delimiters for these strings.
Capability inventory: The membrane CLI can perform network operations and execute generated actions.
Sanitization: Absent. The skill does not specify any sanitization for the input variables.
Note: This represents the standard vulnerability surface for natural-language-to-action tools and is considered low risk in this context.

docmosis