The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze and review user-provided source code, database configurations, and persistence mappings. This task inherently involves ingesting untrusted data that could contain malicious instructions hidden in code comments or metadata.
Ingestion points: Java source files (AggregateRoot, Repository, PO), Maven/Gradle configuration files (POM), and database schemas.
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" wrappers for the agent to use when reading user code.
Capability inventory: The skill is strictly instructional and does not invoke subprocesses, network tools, or file-writing capabilities.
Sanitization: No explicit sanitization or filtering of external code content is required by the instructions.
[SAFE]: The skill demonstrates a strong security posture by instructing the agent to ensure test data is masked and that logs do not expose credentials or private payloads. It also correctly advises users to manage secrets via .env files and refuses to accept H2 unit tests as production-grade proof for concurrency or transaction safety, preventing potential misconfigurations in live environments.

ddd4j-data