optim-agent
LLM agents as your hyperparameter optimizer.
Optim-agent is an agent skill that enables LLM-powered coding agents to perform automated hyperparameter optimization and parameter tuning. Rather than treating configuration spaces as anonymous coordinates, it lets agents like Claude Code, Codex, and OpenCode reason semantically about parameter meanings, study trial history, and propose informed next configurations. The skill is valuable when evaluations are expensive and data budgets are too small for classical surrogate models to excel. Developers integrate it via Python or load it as a skill in agent sessions where the agent reads project code and drives optimization loops directly.
Key Features
Use Cases
- 01Tuning neural network training hyperparameters such as learning rates, batch sizes, and regularization settings
- 02Optimizing inference configurations including quantization levels, batching strategies, and decoding parameters
- 03Tuning reinforcement learning system parameters like exploration schedules and policy thresholds
- 04Optimizing quantitative research signals with parameters for thresholds, windows, and rebalancing rules
- 05Configuring gradient boosting classifiers for credit default prediction or other tabular tasks
- 06Tuning scientific simulation inputs and solver settings to minimize error or runtime
Related Skills
View moresuperpowers
An agentic skills framework & software development methodology that works.
skills
Skills for Real Engineers. Straight from my .agents directory.
skills
Public repository for Agent Skills
ponytail
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
optim-agent — FAQ
What is optim-agent?+
Optim-agent is an agent skill that automates hyperparameter optimization by having LLM coding agents propose parameter configurations based on semantic understanding of the search space and trial history. It combines the meanings of parameters with observed outcomes to suggest the next configuration to evaluate.
How do I install optim-agent?+
Install via PyPI with 'pip install optim-agent' for the Python package, or add it as a skill/plugin to Claude Code, Codex, or OpenCode using their respective installer commands. You need one authenticated agent CLI (claude, codex, or opencode) on your PATH.
Which AI clients work with optim-agent?+
Optim-agent works with Claude Code, Codex, and OpenCode as backend agents. It can also integrate with Claude Desktop and other coding assistants when loaded as a skill that the agent reads and executes.
Do I need API keys to use optim-agent?+
Yes, you need an authenticated agent CLI on your PATH. For Claude backends you need an Anthropic API key; for paid OpenAI models you need an OpenAI key. Free OpenCode-hosted models require no API key.
Is optim-agent free to use?+
The optim-agent software is MIT licensed and free. However, agent backend calls consume API credits; OpenCode provides free model access while Claude and Codex backends require paid API keys.
What are the prerequisites for optim-agent?+
You need Python 3.x, pip, and one authenticated agent CLI (claude, codex, or opencode) accessible in your PATH. Optional dependencies like ml or rl extras are needed for specific example benchmarks.
How do I install optim-agent?+
Open the source repository on GitHub and follow its README. optim-agent is a skill — MCP Agents Market links you directly to the official repo.
Is optim-agent free?+
optim-agent is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.