Run autonomous modify-commit-verify loops to optimize code metrics in git repositories.
npx skills add https://github.com/quanganh208/cookmate --skill ck-loop-quanganh208
Please help me install this Agent Skill. Skill: ck:loop Source: https://github.com/quanganh208/cookmate/tree/main/.opencode/skills/ck-loop Command: npx skills add https://github.com/quanganh208/cookmate --skill ck-loop-quanganh208
This Skill automates repeated, measurable improvements to code and build metrics (for example test coverage, bundle size, lint error counts, and build time) by running an iterative modify-commit-verify loop that learns from git history and safely reverts regressions.
Start an autonomous loop to increase test coverage in src/utils by specifying a Goal, a Scope glob (for example src/utils/**/*.ts), and a Verify command that prints a single numeric coverage percentage to stdout.
High-intent search queries and answers about installing and using this skill.
Automate metric-driven optimization by running an autonomous modify-commit-verify loop that proposes atomic changes, checks a numeric metric, and reverts regressions in git-tracked repositories.
An autonomous iterative loop proposes one atomic code change per cycle, commits it, verifies a numeric metric, and keeps or reverts the change based on whether the metric improves or regresses.
Yes, you can specify an optional guard command to run regression checks alongside the verify command, ensuring the autonomous loop safely discards changes that fail guards.
Yes, a clean git working tree is required before starting the loop, because the tool commits proposed changes before verification and reverts failures directly within the repository.
The loop requires a verify command that prints a single numeric value within a short timeout, relies on git history for reverts, and may waste cycles if consecutive discards occur.
The loop automatically appends structured iteration rows, including baseline metrics and outcomes, to a loop-results.tsv file for reproducible optimization tracking across the repository.
hotriluanRun iterative experiments to optimize measurable metrics in git repositories.
shishiAutomate iterative code reviews with Codex and Claude Code patches.
massgenRun MassGen experiments and generate ANALYSIS_REPORT.md from traces.
victorzhukProfile and benchmark Go applications to optimize CPU, memory, and concurrency performance.
