integrate-any-external-app

Pass

Audited by Gen Agent Trust Hub on Feb 24, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses npx to fetch and execute the @membranehq/cli package. This is an official utility provided by the skill author (Membrane Inc) and is used for its intended purpose of platform interaction.
  • [COMMAND_EXECUTION]: The skill executes local shell commands via the Membrane CLI to manage authentication, list connections, and run actions. These operations involve accessing the tool's own credentials in ~/.membrane/credentials.json and communicating with the vendor's API at https://api.getmembrane.com.
  • [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes data from external applications (e.g., Slack messages, Jira tickets) via integration actions.
  • Ingestion points: Data returned from external app actions (e.g., action run) and agent session updates (e.g., agent-session get).
  • Boundary markers: None explicitly defined in the provided workflow to distinguish untrusted external data from the agent's instructions.
  • Capability inventory: The skill can execute arbitrary actions on connected external apps and initiate agent sessions that generate integration code.
  • Sanitization: No explicit sanitization or filtering of external application content is described in the workflow.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 24, 2026, 11:26 PM