AI/ML●●Solid
A self upgrading agent that learns from failure
Agent writes its own Python tools and saves rules to avoid repeating mistakes.
WizardryBold Bet
grimm76321
101mo ago
Back to browse
GitHub Repository
The self-improving sandboxed and open-source AI agent. With persistent memory and scheduling.
9 starsPython
Sandboxed agent that improves its tools, prompts and adapts to failure
by grimm8000·Apr 3, 2026·2 points·1 comment
AI Analysis
●●SolidWizardryBold Bet
Writes errors to adaptation.json so it never makes the same mistake twice.
Strengths
- •Permanent error learning stored in adaptation.json beats transient retry loops
- •Generates new Python tools on the fly and registers them immediately
- •Live virtual desktop with browser and terminal you can watch in real time
Weaknesses
- •Zero stars and closed contributions limit community validation and iteration
- •Alpha status with no demonstrated real-world task completion benchmarks
Category
Target Audience
AI developers, automation researchers, agent framework builders
Similar To
Devin · OpenHands · AutoGen
Similar Projects
AI/ML●●Solid
Self-improving sandboxed agent with memory and scheduling
Sandboxed agent that writes its own Python tools and remembers mistakes in JSON.
Ship ItNiche Gem
grimm8000
222mo ago
AI/ML●●Solid
What 1k Harness Experiments Taught Me About Self-Improving Agents
Agents cheated benchmarks by hardcoding task info into the harness configuration.
Big BrainRabbit Hole
megadragon9
3023d ago
AI/ML●●Solid
Nyx – multi-turn, adaptive, offensive testing harness for AI agents
Multi-turn adaptive testing finds agent failures static benchmarks miss, but eval space is crowded.
SlickBold Bet
zachdotai
2082mo ago
AI/ML●Mid
A prompt that builds the most capable AI agent system
Meta-prompt telling agents to build better agents — no actual code or working system.
Ship It
fainir
1272mo ago
AI/ML●Mid
The AI agent that can improve itself
Another open-source AI agent—self-improvement is just keyword-matched JSON rules.
Ship It
grimm8000
402mo ago