$ loading_
Build and iterate meta-harness scaffolding for fixed models via propose-score-Pareto loops.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "harness-forge" yet — see the docs or source repo.
Using the harness-forge approach, design an iterative meta-scaffold for a fixed model: memory, retrieval, context assembly, and prompting strategy, plus a propose→score→Pareto improvement loop.
A practical scaffolding plan and iteration workflow.
Please run propose→score→Pareto evaluation on these 3 harness variants, compare their tradeoffs in accuracy, cost, and context utilization, and recommend which to keep for the next round.
A comparison scorecard, Pareto-optimal options, and next-step recommendations.
I have an existing Claude Code workflow; refactor it into a Meta-Harness style by identifying memory, retrieval, prompting, and feedback parts, then propose the revised module structure.
A module breakdown, responsibility map, and refactoring recommendations.
Scaffold a branded AI agent harness with CLI, MCP, memory, and learning.
Automatically distill, update, and prune reusable skills from real coding sessions.
Build AI agents quickly with a model-driven approach and minimal code.
Expose internal agent tools as a standard MCP server for unified access.
Optimize coding agents with performance, security, memory, and research-first workflows.
Explore Claude Code architecture and learn AI agent harness design principles.