The Agent Skills Directory

[DATA_EXFILTRATION]: An absolute local file path is hardcoded in the reference material, which exposes details of the author's local file system structure.
Evidence: /Users/wandl/workspaces/workspace-ddd4j/workspace-ddd4j-boot/ddd4j-quarkus found in references/source-evidence.md.
[INDIRECT_PROMPT_INJECTION]: The skill processes external project data and configuration files, creating a vulnerability surface if those files contain malicious instructions.
Ingestion points: The skill analyzes source code, directory structures, and Maven POM files (found in SKILL.md and references/workflow.md).
Boundary markers: There are no specific delimiters or instructions provided to differentiate user-provided code from the agent's internal instructions.
Capability inventory: The skill instructs the agent to "Run Arc/QuarkusTest suites" (SKILL.md), which involves executing code via build tools or the shell.
Sanitization: No sanitization or verification of the analyzed code is performed prior to processing or test execution.
[REMOTE_CODE_EXECUTION]: The instructions suggest using npx to fetch and execute additional skill modules from a remote repository.
Evidence: Multiple sections in SKILL.md suggest executing npx skills add full-stack-skills/ddd4j-skills to install related tools.
Note: The referenced package belongs to the same vendor as the skill author ('full-stack-skills'), which aligns with expected vendor functionality.

ddd4j-quarkus-architecture