The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script to coordinate between different storage layers.
Evidence: Calls to memory-manager.py --action get-memory-status and memory-manager.py --action search-local are used to manage memory layer connectivity and local indexing.
[EXTERNAL_DOWNLOADS]: The skill references and connects to well-known technology services for data storage and retrieval.
Evidence: Mentions integration with Pinecone, Qdrant, Graphiti, and Supermemory for vector and knowledge graph search operations.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes data retrieved from external and local databases.
Ingestion points: Results retrieved from vector databases (Pinecone, Qdrant), knowledge graphs (Graphiti), and local file-system indices.
Boundary markers: The process description does not explicitly detail the use of delimiters or 'ignore' instructions for the retrieved data content.
Capability inventory: The agent has capabilities to read local files in ~/.claude-marketing/ and execute the memory-manager.py script.
Sanitization: No explicit sanitization or validation of the retrieved knowledge content is described before presentation to the user or agent.

search-knowledge