The Agent Skills Directory

Data Ingestion and Processing: The skill is designed to ingest and analyze repository contents and security policies. While processing external data is a potential vector for indirect prompt injection, the instructions include specific guardrails to treat such content strictly as data for analysis, preventing it from influencing the agent's operational logic.
Internal Capability Management: The skill manages threat model data using designated reference paths and caching mechanisms. The process involves verification steps for model scope and evidence, ensuring that the generated output remains consistent with the repository's actual architecture.

threat-model