Store and retrieve short-term operational memories as semantic vectors.
npx skills add https://github.com/tao3k/xiuxian-artisan-workshop --skill memory-tao3k
Please help me install this Agent Skill. Skill: memory Source: https://github.com/tao3k/xiuxian-artisan-workshop/tree/main/assets/skills/memory Command: npx skills add https://github.com/tao3k/xiuxian-artisan-workshop --skill memory-tao3k
Enables agents and developers to persist and recall short-term operational knowledge such as transient findings, workarounds, and recent execution context that would otherwise be lost between sessions.
Save a transient operational insight by asking the agent to store a one-sentence finding with metadata like domain and kind so it can be semantically retrieved later.
High-intent search queries and answers about installing and using this skill.
Store short-term operational insights as vector embeddings to persist transient findings, workarounds, and recent execution context for later semantic recall. You save a one-sentence finding with metadata like domain and kind, enabling fast retrieval across agent sessions.
LanceDB serves as the vector store for semantic memory management, persisting temporary operational notes as embeddings. It supports IVF-FLAT index creation and optimization to enable fast semantic search and retrieval of recent execution context during agent workflows.
Semantic search lets you query recalled incident notes and temporary workarounds using embedding-based relevance scores. By storing short-term operational memories as semantic vectors, you can retrieve past execution context even when troubleshooting issues like repeated timeouts.
PyYAML is required to parse and load skill manifests into semantic memory. Once loaded, the system creates or optimizes IVF-FLAT indexes within the LanceDB or Rust-backed vector store and retrieves memory statistics for diagnostics.
Generate embeddings for short-term operational memory by saving transient insights through an MCP-facing API. The system handles embedding generation and persists the vectors to a Rust-backed vector store, making them available for semantic search and recall.
Vector stores for transient operational notes are not suited for permanent knowledge retention or structured relational data. They are designed for short-term operational context, temporary workarounds, and recent execution findings that support semantic search rather than long-term archival storage.
Fudan-SMI-labPreinstall PyYAML before torch-npu to prevent transitive dependency failures.
itallstartedwithaideaIngest private documents into a vector index and generate cited answers.
rukikotooQuery the econometrics RAG knowledge base to retrieve relevant document chunks.
CUexterStore embeddings and metadata in an open-source vector database.
