dev_skill

This skill orchestrates ultra-efficient end-to-end development with parallel tasks and 90% test coverage, delivering rapid, reliable backend routing and
  • Go

2.1k

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill cexll/myclaude --skill dev

  • SKILL.md10.4 KB

Overview

This skill orchestrates an extreme lightweight end-to-end development workflow that enforces early backend selection, targeted requirement clarification, parallel task execution, and mandatory ≥90% unit-test coverage. It coordinates multi-agent code work through a single codeagent-wrapper interface and deterministic backend routing per task type. The flow is seven steps long and is optimized for minimal overhead and fast, auditable results.

How this skill works

The orchestrator begins by asking which AI backends are allowed, then runs a short series of clarification questions and creates a Todo task list. A deep technical analysis is delegated to codeagent-wrapper in analysis mode to explore the codebase and break work into 2–5 parallelizable tasks. All code changes are executed through codeagent-wrapper --parallel with backend routing based on task type, and coverage is validated to meet the 90% requirement before finalizing.

When to use it

  • Implementing small-to-medium features that benefit from parallelized task execution.
  • Refactoring or adding features where test coverage is required to be ≥90%.
  • When you want deterministic backend selection and enforced toolchain usage.
  • Coordinating multi-agent contributions without manual code edits.
  • Projects where fast iteration and clear audit trails are required.

Best practices

  • Always respond to the initial backend-selection prompt; backend choices drive routing for all tasks.
  • Keep clarification answers concise and precise to finish 2–3 Q/A rounds quickly.
  • Design features so they can be split into 2–5 independent tasks with clear file scopes.
  • Prefer codex for default work, gemini for UI, and claude for quick-fix tasks—respect allowed_backends.
  • Ensure tests are written alongside code; aim for incremental coverage targets to reach ≥90%.

Example use cases

  • Add a new API endpoint with accompanying unit tests and integration test harness.
  • Refactor a service module into smaller components while keeping coverage high.
  • Introduce a small frontend component and scss change routed to the UI backend.
  • Apply quick configuration or CI fix using a fast backend for rapid turnaround.

FAQ

If you select only codex, every task—default, ui, or quick-fix—will be routed to codex without exceptions.

Can I skip requirement clarification?

No. The workflow mandates a requirement-clarification step using AskUserQuestion and a TodoWrite before analysis.

How are code changes made?

All code modifications must be performed via codeagent-wrapper --parallel; direct Edit/Write tools are never used.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational