neutron-agent-memory
от naeemmaliki036
Search & ResearchКратко
Store and retrieve agent memory using Neutron API.
Описание
Что делает этот навык
Provides persistent memory storage for AI agents using the Neutron API. Saves text as searchable seeds and manages agent context objects across sessions. Supports semantic search to retrieve relevant past content by meaning rather than exact match.
Semantic search over stored seeds lets agents find relevant past context without needing exact keyword matches, which plain key-value stores cannot do.
Когда его использовать
- Remembering user preferences across separate agent sessions
- Finding past conversation snippets by semantic similarity
- Tracking episodic agent context for long-running workflows
- Storing structured metadata alongside free-text memories
- Persisting full conversation thread snapshots for later retrieval
Пример сценария
User asks: 'What did we discuss about blockchain last month?'
- 1
Run semantic search against stored seeds with query 'blockchain'
- 2
Retrieve top matching seeds above similarity threshold
- 3
Pull associated agent context for session metadata
- 4
Combine seed content and context into a coherent summary
- 5
Return the relevant past discussion to the user
Agent surfaces the closest matching conversation snapshots with blockchain content, ranked by semantic relevance
Требования
- NEUTRON_API_KEY environment variable
- NEUTRON_APP_ID environment variable
- Neutron account with API credentials
Установка
clawhub install naeemmaliki036/neutron-agent-memoryВыполните команду CLI или вставьте промпт в Claude Code, Cursor, Codex — любой кодинг-агент, которым пользуетесь.