ddd4j-cloud-release
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes project-specific data including Maven POM files, Git SHAs, and remote repository metadata to automate release tasks. This represents a standard attack surface for indirect prompt injection.
- Ingestion points: The skill reads local POM files, source code, and remote repository metadata (references/source-evidence.md, references/workflow.md).
- Boundary markers: The skill does not use explicit boundary delimiters but follows a rigid, step-by-step workflow to verify the "source of truth."
- Capability inventory: Filesystem access, Git operations, and execution of Maven build commands.
- Sanitization: The instructions include mandatory rules to redact all credentials and sensitive connection strings from output logs.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute Maven commands (e.g.,
-Ufor clean-cache verification), Git commands to record history, and compatibility verification scripts. These are standard operations for artifact release management. - [EXTERNAL_DOWNLOADS]: The skill includes instructions for the user or agent to install related tools using
npx skills add. These references target the vendor's own repository (full-stack-skills/ddd4j-skills) and are standard for the skill platform's ecosystem.
Audit Metadata