Skip to content

CHANGELOG.md

Latest commit

 

History

History
141 lines (107 loc) · 6.36 KB

CHANGELOG.md

File metadata and controls

141 lines (107 loc) · 6.36 KB
Raw

Changelog

All notable changes to the StoryCore Gem Protocol will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[1.0.0] — 2026-03-04

🎉 First Stable Release — Gem Protocol MVP

Core Gem Engine

  • UserWallet model with balance, tier, and total earned tracking
  • GemTransaction model for full audit trail of all Gem flows
  • GemEscrow model for P2P compute payment escrow (atomic delivery)
  • WorkerNode model for GPU compute provider registration
  • TaskCategory model for compute task types and VRAM requirements
  • GemEngineStandalone engine: process_transaction, transfer_gems, create_escrow, release_escrow, cancel_escrow

REST API Endpoints (/v1/gems/)

  • GET /balance/{user_id} — Query wallet balance and tier
  • GET /history/{user_id} — Full transaction history
  • POST /transfer — P2P Gem transfer between users
  • POST /escrow/create — Create atomic compute escrow
  • POST /escrow/release/{escrow_id} — Release Gems to worker on task completion
  • POST /escrow/cancel/{escrow_id} — Cancel and refund escrow to sender
  • POST /worker/register — Register GPU node with vRAM and capabilities
  • POST /worker/heartbeat/{worker_id} — Node presence keepalive
  • GET /workers — List all registered compute nodes
  • GET /task-categories — List available compute task types with costs
  • POST /tasks/seed — Seed default categories: video, audio, science, research

Human Contribution Analysis (Pillar 1)

  • POST /ai/analyze-contribution — AI-scored effort quantification
    • Heuristic scoring: effort × impact multiplier (keywords: fix, critical, security)
    • Safety cap: max 100 Gems per contribution

AI Agent Economy (Pillar 4)

  • POST /ai/agent/register — Register AI agents (OpenClaw, LangChain, AutoGPT, custom)
    • Open mode (no signature) and HMAC-SHA256 secured mode
  • POST /ai/agent-contribution — Submit agent output for Gem valuation
    • Formula: base_value × quality_score × novelty_multiplier × reputation_factor
    • Task types: research (15), code (25), creative (12), summarization (10), orchestration (30)
    • Novelty deduplication via SHA-256 content hashing (duplicate penalty: 0.1×)
    • Logarithmic reputation growth with contribution history (max factor: 3.0)
    • Rate limiting: 50 contributions/hour per agent_id
    • Contributions >50 Gems flagged for human review; hard cap at 150 Gems
  • GET /ai/agents — Agent leaderboard sorted by total Gems earned

Infrastructure

  • Async SQLAlchemy with aiosqlite for zero-config local development
  • FastAPI with full OpenAPI/Swagger docs at /docs
  • Background task for stale worker cleanup
  • Hot-reload development server via uvicorn --reload

Documentation

  • README.md — Full Gem Protocol vision: 5 pillars documented
    • Pillar 1: Human Contribution (GitHub integration)
    • Pillar 2: Compute Sharing (P2P Mesh + Escrow)
    • Pillar 3: Universal Utility (Science, Research, Agents)
    • Pillar 4: AI Agent Economy (Proof-of-Intelligence, OpenClaw bridge)
    • Pillar 5: API Budget Recycling (Zero-Waste Intelligence)
  • LICENSE — MIT License
  • CHANGELOG.md — This file

Architecture Notes

  • In-memory rate limiter and agent registry (MVP) — Redis recommended for production
  • Heuristic scoring engine (MVP) — LLM-based evaluation planned for v1.1
  • HMAC-SHA256 optional in open/dev mode; mandatory when shared_secret is configured

[1.1.0] — 2026-05-05

🎉 Added: Frontend Dashboard (React + TypeScript + Tailwind)

  • Complete React SPA with Vite build tooling
  • Tailwind CSS v4 with dark/light theme support
  • Shadcn-style UI components (Button, Card, Badge, Input, Select, Label)
  • 6 pages: Dashboard, Applications, AI Agents, Escrows & Transfers, Workers, Wallet
  • React Query (@tanstack/react-query) for API caching and state management
  • Proxy configuration: /v1/* → localhost:8001 in dev mode
  • TypeScript types matching all Pydantic API models
  • Loading, empty, and error states for all pages
  • Sidebar navigation with routing (react-router-dom)

New Pages

Page Features
Dashboard Service status, agent leaderboard, recent escrows
Applications Register apps with reward rules, list agents by app
AI Agents Agent list with reputation, register agent form, contribution scoring
Escrows Escrow list with release/cancel actions, create escrow, P2P transfer
Workers Worker node list, register worker, task categories
Wallet Balance/tier lookup, transaction history

🔒 Security Fixes

  • FastAPI upgraded 0.115.0 → 0.115.14 (Starlette 0.46.2, fixes CVE-2024-47874 DoS)
  • PyJWT upgraded 2.10.1 → 2.12.1 (fixes CVE-2024-53862 crit header)
  • python-dotenv upgraded 1.0.1 → 1.2.2 (fixes CVE-2024-48879 symlink)
  • Mako pinned >=1.3.7 (fixes path traversal CVE)
  • Alembic upgraded 1.13.1 → 1.18.4 (pulls fixed Mako)
  • Starlette multipart DoS (moderate) patched via upgrade to 0.46.2

🐛 Bug Fixes

  • setup_app.py: relative imports replaced with absolute imports (crash fix)
  • simulate_cine_gem_flow.py: removed broken imports to backend.* (ModuleNotFoundError)
  • api/gems.py: removed unused func import
  • main.py: removed unused imports (Depends, HTTPException, Header, Request)
  • engine.py calculate_reward: clarified default logic (no longer returns 1 when no __default__ key exists)
  • app/core/engine.py calculate_reward: added max(0, ...) guard (prevent negative rewards)
  • engine.py cancel_escrow: added warning log when sender wallet not found
  • main.py: merged duplicate @app.on_event("startup") handlers

📝 Documentation

  • README updated with frontend setup instructions, badges, and production build guide
  • Fix corrupted emoji character in README

[Unreleased] — Planned for v1.2.0

Planned

  • LLM-based quality scoring (replace heuristics with Gemini/Claude API call)
  • Redis-backed rate limiting and persistent agent registry
  • Agent task queue: StoryCore broadcasts pending tasks; agents claim and process
  • API Budget Recycler: end-of-month surplus detection and auto-routing
  • Multi-signature escrow for high-value compute tasks (>500 Gems)
  • REST webhook callbacks when tasks are assigned to agents