verification-gates
Enforces build, test, and lint gates with a self-correction loop and retry budget.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
القائمة
Enforces build, test, and lint gates with a self-correction loop and retry budget.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Scans the codebase for tech debt — balls of mud, shallow modules, and DRY violations — then presents prioritized bottlenecks and hands off to TDD-driven refactoring via superpowers.
Forces DDD ubiquitous language definition before writing complex logic. Grills the user on business terminology, extracts a glossary, and documents it before handing off to brainstorming.
Sets up a new project for the AI workflow — creates AGENTS.md, mise.toml tasks, initializes CodeGraph, and seeds Engram memory. Optional/auxiliary skill.
Evidence-first development for high-stakes code. Writes a concrete spec, enforces strict TDD, then runs code through a gauntlet of 6 checks and generates an EVIDENCE.md report with raw numbers.
Enforces Red-Green-Refactor Test Driven Development workflow.
Manages Architecture Decision Records (ADRs) when modifying architecturally significant files.
| name | verification-gates |
| description | Enforces build, test, and lint gates with a self-correction loop and retry budget. |
To ensure repository stability, all major changes must pass through a series of "Verification Gates."
npm run build (or equivalent).npm test (or equivalent).npm run lint or ruff check.Whenever a gate fails (e.g., via the gate tool), follow this loop:
stderr and stdout provided by the gate tool.gate tool with the same command.Always run the gate tool for any verification step. If a gate fails, announce "SELF-CORRECTION LOOP START [Attempt X/5]".