package-automq-green

Pass

Audited by Gen Agent Trust Hub on Sep 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATIONDYNAMIC_EXECUTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The green launcher script uses babashka.deps/add-deps to dynamically load and execute Clojure code from the vendor's GitHub repositories at runtime. The dependencies for automq, colors-compute, and other internal libraries are pinned by specific Git SHAs.
  • [EXTERNAL_DOWNLOADS]: The skill fetches multiple implementation libraries from GitHub repositories belonging to the vendor organization. Additionally, it mentions installing third-party tools such as Docker, ACME clients (lego), and managing APT repositories on the provisioned target nodes.
  • [COMMAND_EXECUTION]: The skill provides lifecycle commands (build, create, delete, validate) that execute provisioning workflows involving OpenTofu and Ansible to manage cloud resources.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests configuration from colors.yml, creating a surface for potential injection if the file is sourced from untrusted PRs or environments.
  • Ingestion points: colors.yml is the primary file for desired state ingestion.
  • Boundary markers: None identified in the prompt instructions.
  • Capability inventory: The launcher script and its underlying libraries perform network operations, execute subprocesses (OpenTofu, Ansible), and write local files.
  • Sanitization: A validate command is provided to check the desired state configuration.
  • [PRIVILEGE_ESCALATION]: The provisioning process performs administrative tasks on target hosts, including modifying firewall rules (iptables), managing SSH authorized keys, and updating system repository configurations in /etc/apt/sources.list.d/ubuntu.sources.
  • [DYNAMIC_EXECUTION]: The green script dynamically resolves and loads Clojure namespaces from computed Git paths at runtime, which is standard for Babashka tools but represents dynamic execution of external code.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 13, 2026, 08:44 AM