The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted content from the repository it audits, such as AGENTS.md, CLAUDE.md, and documentation within the .specify/ or openspec/ directories.
Ingestion points: Files like AGENTS.md, CLAUDE.md, and POM files are read to determine the state of the project and its release readiness.
Capability inventory: The skill performs high-privilege operations including git push and mvn deploy to authorized private repositories.
Boundary markers: The instructions lack requirements for using boundary markers or delimiters when the agent interprets these external project files.
Sanitization: While the skill instructs the agent to redact credentials in logs, it does not define input sanitization for instructions found in project documents.
[COMMAND_EXECUTION]: The skill facilitates project lifecycle management by executing shell commands for version control (git status, git push, git worktree) and build processes (mvn deploy, mvn test). These tools are appropriate for the skill's purpose but represent a sensitive capability surface.
[EXTERNAL_DOWNLOADS]: The skill documentation suggests the installation of additional tools and skills from the author's ecosystem, such as full-stack-skills/ddd4j-skills, via the npx skills add command. These are standard vendor-provided resources.

ddd4j-javalin-production-hardening