opentofu-coder

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
  • EXTERNAL_DOWNLOADS (HIGH): The file references/post-provisioning.md includes a bash script template that executes curl -fsSL https://get.docker.com | sh. Piped remote execution (curl to shell) is a high-risk pattern that executes unverified code from external sources, bypassing package manager security controls.\n- COMMAND_EXECUTION (MEDIUM): The automation logic in references/makefile-automation.md utilizes the $(shell ...) function and invokes several system binaries including tofu, jq, and op. These allow the skill to execute arbitrary commands on the host system as part of the automation workflow.\n- DATA_EXFILTRATION (LOW): The Makefile in references/makefile-automation.md performs a network request to ifconfig.me to retrieve the local public IP address. While used for legitimate firewall whitelisting, this involves an outbound connection to an external third-party service.\n- PROMPT_INJECTION (LOW): The skill templates for HCL and Shell scripts ingest external variables (like server IPs and stack names) that could be manipulated. \n
  • Ingestion points: Variable interpolation in references/post-provisioning.md and CLI arguments in references/makefile-automation.md.\n
  • Boundary markers: Absent.\n
  • Capability inventory: Includes SSH execution, OpenTofu resource management, and network operations.\n
  • Sanitization: Relies on standard shell quoting which may be insufficient against sophisticated injection.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:23 PM