Dependency Cache Optimizer

Goal

Improve cache hit rates

How to Run

This loop optimizes CI cache performance by analyzing metrics and iteratively improving dependency caching strategies.

1. 01

   Open your coding agent environment

   Launch Cursor, Claude Code, Codex, OpenCode, or Gemini CLI in your project directory.

2. 02

   Run the kickoff prompt

   Paste the kickoff prompt to initialize the loop and begin cache optimization.

3. 03

   Review agent suggestions

   Examine proposed cache improvements and approve changes before they are applied.

4. 04

   Monitor iterations

   Let the agent run automatically, checking cache metrics after each iteration until the target hit rate is achieved or max iterations reached.

Workflow Steps

1. 01

   Analyze current cache performance

   Review cache metrics to identify packages, configurations, and patterns causing low hit rates.

2. 02

   Identify optimization opportunities

   Detect frequently changing dependencies, missing cache keys, or suboptimal restore strategies.

3. 03

   Propose cache configuration changes

   Generate improvements to cache key structure, dependency pinning, or restoration logic.

4. 04

   Apply and validate changes

   Update cache configuration and ensure changes align with project constraints.

5. 05

   Verify cache hit rate improvement

   Execute check command to measure impact of applied optimizations.

Start the "Dependency Cache Optimizer" loop.

Goal: Improve cache hit rates
Max iterations: 10
Between iterations run: cache metrics
Exit when: Cache hit threshold reached


Act as a CI performance optimization specialist. Your goal is to improve cache hit rates. Run 'cache metrics' to analyze current cache performance. Iteratively identify bottlenecks in dependency caching, propose configuration improvements, and validate changes. Stop when cache hit threshold is reached or after 10 iterations. Always prioritize build stability over aggressive optimizations. Log all changes and ensure check command provides measurable results.

Self-pace this loop. After each iteration, run `cache metrics` and evaluate the output, and only continue if the exit condition is not met (Cache hit threshold reached). Stop when the exit condition passes or 10 iterations are reached. Give a short status update each pass.