The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references installing the neo4j-agent-memory package from the official PyPI registry and using uvx to execute vendor-provided tools. These resources are hosted on well-known services and are consistent with standard development practices.
[COMMAND_EXECUTION]: Documentation includes instructions for pip install, uvx, and registration commands for MCP servers like claude mcp add. These are intended for environment setup and tool configuration, posing no security risk in the context of this reference material.
[PROMPT_INJECTION]: The skill describes a memory system that ingests external user data, creating a potential surface for indirect prompt injection where untrusted content could influence agent context.
Ingestion points: Data enters the system through MemoryClient.short_term.add_message and the memory_store_message tool.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are documented in the provided reference snippets.
Capability inventory: Tools such as memory_search and memory_get_context retrieve the ingested data to build LLM prompts.
Sanitization: The documentation does not specify sanitization or validation logic for the content stored in the graph memory.

neo4j-agent-memory