The Agent Skills Directory

[SAFE]: The skill serves as a legitimate administrative guide for the Wardx server platform. It defines strict boundaries for production access, local database usage, and secret management, explicitly prohibiting the exposure of raw credentials or keys.- [COMMAND_EXECUTION]: The file references/package.md lists standard Node.js development scripts (e.g., npm run lint, npm run test:js, npm run typecheck) for package verification. These commands are documentation for standard development workflows and do not constitute malicious execution or injection.- [INDIRECT_PROMPT_INJECTION]: The skill interacts with external data sources such as logs and project metadata via tools defined in references/tools.md. While these ingestion points exist, the instructions in SKILL.md provide clear guardrails, ordering the agent to never invent data and requiring specific versioning/reasoning for all mutations, which effectively mitigates typical injection risks.

wardx-server