sleepy
51123212c4
Harnesses under analysis: - opencode (Go-based coding agent) - pi (minimal terminal coding harness by Mario Zechner) - hermes (Nous Research agent) - forgecode (AI pair programmer with sub-agents) Each harness folder contains: - repo/: Source code from respective repositories - feedback/localllm/: Community feedback for local/smaller models - feedback/frontier/: Community feedback for frontier models Research focus: Tool handling, skills systems, prompt engineering, context management, and best practices for smaller/local models.
1.4 KiB
1.4 KiB
AGENTS.md
Research/Analysis Folder for opencode
This is the research and analysis folder for the opencode coding harness.
Folder Structure
opencode/
repo/ - opencode-ai/opencode source code
feedback/
localllm/ - Community feedback and performance data for local models
frontier/ - Community feedback and performance data for frontier models
What's Inside
- repo/: The official opencode repository (Go-based coding agent)
- feedback/localllm/: Feedback, benchmark results, and observations from using opencode with smaller/local LLMs
- feedback/frontier/: Feedback, benchmark results, and observations from using opencode with frontier models
Feedback Format
Each feedback file should include:
- Model used (name, size, provider)
- Benchmark results or task performance
- Issues encountered
- What worked well
- Source reference: URL or site where the feedback came from (community posts, Discord, GitHub issues, etc.)
Research Focus
This folder collects data on:
- Tool handling and capabilities
- Skills system effectiveness
- Prompt engineering strategies
- Context management
- Performance on benchmarks (terminal-bench, etc.)
Goal
Extract best practices specifically for smaller/local models and document what works vs. what doesn't for the opencode harness. General use / use with frontier models information should be put in the feedback/frontier folder.