alibabacloud-ecs-code-deploy

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The scripts/deploy_toolkit.py script executes shell commands via the subprocess module to interact with the Alibaba Cloud CLI (aliyun), manage environment virtualenvs, and perform loopback probes on the target server. The tool implements internal regex validation for deployment parameters to mitigate injection risks.
  • [EXTERNAL_DOWNLOADS]: The skill downloads the official Alibaba Cloud CLI and appmanager-cli plugin from the vendor's official OSS buckets (aliyun-cli.oss-cn-hangzhou.aliyuncs.com). It also facilitates the download of standard development runtimes (Node.js, Go, PHP) from well-known repositories during the deployment phase on the target server.
  • [REMOTE_CODE_EXECUTION]: The deployment templates utilize shell piping patterns (curl | bash, curl | php) to install development environments on the target ECS instance. These patterns target established domains like NodeSource and Composer. The skill proactively includes SHA256 checksum and cryptographic signature verification steps for Go and Composer to mitigate supply chain risks.
  • [DYNAMIC_EXECUTION]: The agent dynamically generates shell scripts (start and stop) based on its analysis of the user's project files. These scripts are written to a YAML configuration file (.appmanager/config.yaml) and later executed on remote Alibaba Cloud instances via the appmanager orchestration service.
  • [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface for indirect prompt injection as it reads untrusted content from project files (e.g., README.md, package.json, pyproject.toml) to determine build and run commands. The capability inventory includes network operations and file writes; however, the impact is limited by the isolation of the target ECS instance and the requirement for user confirmation of costs and risks before deployment.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 07:04 PM