coding-quality-loop

Solid

Use when a coding agent must turn a software goal, bug, issue, or refactor into a small, verified, independently reviewed code change.

Code & Development 5 stars 0 forks Updated 3 days ago MIT

Install

View on GitHub

Quality Score: 78/100

Stars 20%
26
Recency 20%
100
Frontmatter 20%
70
Documentation 15%
100
Issue Health 10%
50
License 10%
100
Description 5%
100

Skill Content

# Coding Quality Loop Goal: the smallest correct change, with evidence a human can trust, review, revert, or merge. This file is the only always-loaded text. Load `references/` only when needed. Workers never load either. ## Orchestrator Layer The main session is the **orchestrator**. It thinks hard and makes every decision; workers only execute. - Orchestrator owns: task class, context map, contract, right-size rung, plan, model routing, verdict on findings, stop-if-unsafe. - Workers (implementer, reviewer) receive a **brief, not context**: goal, contract slice, files, commands, done-check. One screen max. No skill text, no references, no repository tour. - Delegate only to frontier models on Claude Code or Codex (see Routing). Droid is a supported install target outside the routed kernel; delegation to Droid is allowed only when an explicit harness routes it. The reviewer must be fresh context and cannot be the implementer. Different-family review is the portable default; an explicit higher-level harness may pin the same model across separate hosts/sessions when deterministic gates and supervisor verification retain final truth. ## Task Class (pick the smallest safe) - **tiny** — typo, one-liner. Edit, run the smallest check. No artifacts. - **small** — local bug, one module. Mini map, minimal fix, targeted test. - **medium** — multi-file, feature, migration, auth/payment/data risk. Contract, plan, right-size gate, independent review, completion record. - **mission** ...

Details

Author
zaingz
Repository
zaingz/coding-quality-loop
Created
1 months ago
Last Updated
3 days ago
Language
Python
License
MIT

Integrates with

Similar Skills

Semantically similar based on skill content — not just same category