AI agent runtime with persistent identity, hybrid long-term memory, time-series forecasting, sandbox-isolated tool calls, 184-persona multi-agent swarm, and a chain-hashed deterministic event log for replay, fork, and audit. Built for finance workflows, daily operations, research, and continuous automation. AGPL-3.0 + commercial. Pre-alpha.
npx add-skill Arch1eSUN/MoiraiAI agent runtime with persistent identity, hybrid long-term memory, time-series forecasting, sandbox-isolated tool calls, 184-persona multi-agent swarm, and a chain-hashed deterministic event log for replay, fork, and audit. Built for finance workflows, daily operations, research, and continuous automation. AGPL-3.0 + commercial. Pre-alpha.
BenchClaw is a Codex/OpenCode skill workflow for benchmark construction, evaluation, and maintenance. It standardizes the full pipeline—from idea drafting and data generation to evaluation, reporting, failure diagnosis, and skill refinement—so agents can build reproducible, auditable benchmarks with clear quality gates, lineage, and rollback.
MCP-based equity research agent — three FastMCP servers, Claude Code skills, and a standalone CLI mode that produce 8-section interactive HTML research reports.
Deliver and enforce engineering discipline with 38 proven skills and a CLI designed for real-world multi-agent CI/CD workflows.
Claude Code skill: Export Mac WeChat chat to Obsidian with image decryption
One stock code → full company story
A collection of lightweight, stdlib-only Python utilities built for AI agents