directus

Pass

Audited by Gen Agent Trust Hub on Apr 24, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to install the @membranehq/cli package globally via npm. This is a known tool provided by the skill's author (membranedev) to facilitate platform integrations.
  • [COMMAND_EXECUTION]: The skill utilizes shell commands via the membrane CLI to manage connections, discover actions, and execute requests against the Directus API. These commands are necessary for the skill's primary functionality.
  • [PROMPT_INJECTION]: The skill retrieves and processes external data from Directus, which creates a surface for indirect prompt injection where malicious content in the database could attempt to influence agent behavior.
  • Ingestion points: Data is retrieved from Directus through membrane action run (e.g., list-items) and membrane request commands.
  • Boundary markers: The instructions do not specify any markers or delimiters to isolate untrusted data from the system prompt.
  • Capability inventory: The skill allows write operations (create/update items) and arbitrary API requests via the proxy command.
  • Sanitization: No explicit validation or escaping of the retrieved content is implemented before it enters the agent's context.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 24, 2026, 04:28 PM