twingate-terraform

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches Terraform providers and reference scripts from the official Terraform Registry and Twingate-Solutions GitHub repositories.
  • [REMOTE_CODE_EXECUTION]: Documentation for GCP deployments includes a pattern that downloads and executes the official Twingate Connector setup script via a piped shell command (curl | bash) from binaries.twingate.com.
  • [INDIRECT_PROMPT_INJECTION]: The skill is instructed to read existing .tf files in the user's repository to maintain configuration consistency, which introduces a theoretical surface for malicious local content to influence agent output.
  • Ingestion points: Local .tf files (referenced in SKILL.md globbing instructions).
  • Boundary markers: None specified for the file ingestion process.
  • Capability inventory: Shell command execution via terraform, make, and git.
  • Sanitization: None specified for ingested file content.
  • [COMMAND_EXECUTION]: Provides instructions for using standard development and infrastructure tools, including git, make, and terraform, for provider testing and resource provisioning.
  • [SAFE]: The skill actively promotes security by instructing the agent to avoid hardcoding API tokens and to mark connector tokens as sensitive in Terraform configurations to prevent data exposure.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 09:20 PM
Security Audit — agent-trust-hub — twingate-terraform