Menu
RAG-augmented AI assistant powered by Google Gemini with MongoDB-backed knowledge base, chat history, and feedback learning.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
| name | assistant |
| description | RAG-augmented AI assistant powered by Google Gemini with MongoDB-backed knowledge base, chat history, and feedback learning. |
| metadata | {"amberclaw":{"emoji":"🤖","requires":{"env":["GEMINI_API_KEY"],"services":["mongodb"]}}} |
RAG-augmented conversational AI integrated into AmberClaw as native tools.
assistant_chatChat with Personal Assistant using context from conversation history, knowledge base, and feedback.
Use the assistant_chat tool to ask questions, get help, or have a conversation.
Personal Assistant remembers previous messages within a session and searches its knowledge base automatically.
Parameters:
message (required) — The message to sendsession_id (optional) — Session ID for context continuity (defaults to today's date)assistant_knowledge_searchSearch the RAG knowledge base built from Google Drive documents.
Use assistant_knowledge_search to find specific facts, documentation, or reference material
from the user's uploaded knowledge base.
Parameters:
query (required) — Search querytop_k (optional) — Number of results (1-10, default: 3)Set in ~/.amberclaw/config.json under assistant:
{
"assistant": {
"enabled": true,
"model": "gemini-2.5-flash",
"mongodbUri": "mongodb://localhost:27017",
"googleApiKey": ""
}
}
Environment variables: GEMINI_API_KEY or GOOGLE_API_KEY, MONGODB_URI
amberclaw assistant telegram — Start Telegram botamberclaw assistant sync — Sync Google Drive → RAG knowledge baseamberclaw assistant status — Check connection statusCreate or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Codebase mapping and architectural maintenance. Use to prevent architectural drift.
Master behavior injection for full Claude Code v2.1.110 emulation. Use when needing peak performance.
Plan-Explore-Task-Verify workflow for massive tasks. Use for full feature implementation.
Strict safety rule: read before write. Use when working on complex or sensitive codebases.
Locks responses to Claude Code CLI style. Use to ensure clean and uniform output.