The Agent Skills Directory

[SAFE]: The skill provides detailed architectural instructions for an AI agent acting as a senior caching architect. There are no signs of prompt injection, data exfiltration, or obfuscation. The recommendations follow industry standards and include specific security warnings against using unsafe serialization formats like Python pickle or Java Serialization for distributed caches (references/serialization-memory.md).
[INDIRECT_PROMPT_INJECTION]: The skill outlines a process for analyzing external data such as system requirements and production access logs (references/requirements.md, references/hot-keys.md). While this represents a data ingestion surface common to analytical agents, it does not include high-risk capabilities like automatic script execution based on that data.
Ingestion points: System descriptions, performance metrics, and production access logs.
Boundary markers: The instructions do not define specific delimiters for separating user-provided logs from the agent's internal instructions.
Capability inventory: The skill is designed for producing architectural deliverables and configuration recommendations rather than executing code directly on the host system.
Sanitization: As a design framework, the skill focuses on human-in-the-loop architectural decisions rather than automated data processing, reducing the risk of unintended command execution.

caching