Find the input burden in an elaborate personal-agent system
At a glance
A founder built MUSASHI, then found that daily manual reporting made his own assistant difficult to keep using. This workflow is listed under personal-agents, agent-memory, usability, failure-case.
Agents / tools
MUSASHI, Python, FastAPI, LangGraph, Supabase, Pinecone, Gemini, Claude Code
The private system assigns nine agent roles to dialogue, weekly review, pattern detection, strategy, information collection and supporting functions. It uses Python, FastAPI and LangGraph with Supabase and Pinecone, and records disagreements and council objections.
How it went
During daily use, the founder found that reporting events and correcting misunderstandings demanded too much effort.
Worth knowing
Automatic collection from calendars, email and other sources was a proposed direction for a future public version, rather than an established outcome of this private build.
Save ideas in this browser. Upvotes count across the board.
Indie Hackers founder Sean Prentice used Claude Code to build an orchestrator + marketing director + specialist agents (content/ads/research) for his SaaS; outputs land in Slack and nothing posts or spends without his explicit say (~$10/mo self-use).
MIT Python package signs every policy-gated tool call into a hash-chained .tkb file verifiable offline (Python+TS); works across Claude Code, OpenAI Agents, LangChain/LangGraph, MCP and more—agents can edit their own logs but not hide edits when the signer is out of reach.