from claude-skill-registry357
Agent orchestration library for conversational AI — coordinates LLM completions, memory, tool execution, and multi-turn flows; useful for building chat agents a
libagent provides an orchestration layer for conversational agents, exposing primitives like AgentMind for reasoning and AgentAction for tool execution. It coordinates LLM calls, memory clients, and tool clients to manage multi-turn dialogues and streaming responses.
Use libagent when building conversational assistants that require tool integration, streaming outputs, or retrieval-augmented generation (RAG) flows. It's appropriate for developers wiring memory stores, RPC tool clients, or custom toolchains into an agent.
Suited to developer-facing assistant agents that can execute JavaScript/Node workflows and integrate with memory/llm clients (e.g., Claude Code, local Node-based agent runtimes).
libagent is a documentation-only skill describing an AgentMind/AgentAction orchestration library for conversational AI. It provides usage patterns as code snippets but contains no executable scripts. The SKILL.md is thin — mostly API usage examples tied to a specific @copilot-ld/libagent npm package, with no setup instructions, error handling guidance, or real agent workflow logic. Essentially a library README disguised as a skill.
Documentation-only skill with no executable content. The skill describes a library (@copilot-ld/libagent) rather than providing an agent capability itself. Low usefulness as it doesn't enable any autonomous action — it's just API docs for a JS library. Architecture is flat with no scripts/, references/, or separation of concerns. Code quality is low because the SKILL.md lacks specificity in triggers, has no error handling guidance, and reads like a library README rather than a skill.
