experiment-loop
Experiment Loop Pattern
Overview
The experiment loop is a systematic approach to iterative code improvement where each change is measured against a baseline metric and kept only if it improves the result. Inspired by karpathy/autoresearch and generalized for any measurable code quality metric.
The Core Pattern
1. DEFINE → Choose a measurable metric and set constraints
2. BASELINE → Measure the current state
3. MODIFY → Make one targeted change
4. MEASURE → Re-evaluate the metric
5. DECIDE → Keep if improved, revert if not
6. LOG → Record what was tried and the outcome
7. REPEAT → Go to step 3 (until done or plateau)
When to Use This Pattern
More from pfangueiro/claude-code-agents
deep-read
Comprehensive codebase reading engine. Systematically reads actual source code line by line through a 6-phase protocol — scoping, structural mapping, execution tracing, deep reading, pattern synthesis, and structured reporting. Source code is the source of truth. Use when needing to truly understand how code works, not just what documentation claims.
47git-workflow
Git workflow best practices and patterns. Use this skill when working with git operations, creating commits, managing branches, handling pull requests, or establishing team git workflows. Provides guidance on commit messages, branching strategies, and collaboration patterns.
11ci-cd-templates
Production-ready CI/CD pipeline templates for GitHub Actions, GitLab CI, and CircleCI
7docker-deployment
Production-ready Docker configurations, multi-stage builds, and deployment best practices
7execute
Orchestrated task execution engine. Decomposes any goal into small atomic tasks, plans dependencies, selects the right agent/tool/MCP server for each, executes in optimally parallel batches, and tracks everything. Use when given a complex, multi-step goal that benefits from structured decomposition and full tool utilization.
6library-docs
Quick access to up-to-date library documentation using MCP. Use this skill when you need to reference official documentation for libraries, frameworks, or APIs. Leverages the context7 MCP server to fetch current docs for React, Next.js, Vue, MongoDB, Supabase, and hundreds of other libraries. Complements the documentation-maintainer agent.
6