luisnquin/nixos-config/dev-flow

Claude Code and agent skill by luisnquin/nixos-config

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Default loop for a bug or change. Scope to what was asked, reproduce on the real target, fix the cause, instrument instead of guessing after two failed attempts, prove with before/after evidence, then deliver only as far as allowed. Use for any bug fix or feature implementation.

Install

npx vskill@latest install luisnquin/nixos-config/dev-flow

SKILL.md

Dev flow

The default loop for a bug or a change: scope, reproduce, fix, prove, deliver. Do not skip a step. Do not take one the user did not allow.

1. Scope

  • Restate what was asked in one line. Do only that. A related problem you notice

is a line in the report, not an edit.

  • A question ("wdyt", "what's next", "¿tiene sentido?") gets an answer, not an

implementation.

  • Before writing new code, find the repo's existing pattern, component, hook or

helper that does the same thing, and follow it.

  • Note what is allowed for this task: build, eval, tests, dev server, device,

deploy, commit. "Don't build" and "no tests" hold until the user lifts them.

2. Reproduce

  • Trigger the bug on the real target before touching code: a test, the CLI, or

a device through phone. Write down the exact steps and the observed output.

  • If it does not reproduce, say so and stop. Never fix blind.

3. Fix

  • Make the smallest change that removes the cause, not the symptom.
  • No hardcoding to make one case pass. No new abstraction unless asked.
  • Failed twice ("still the same", "sigue igual", "nada"): stop guessing.
  1. Add instrumentation (logs, a debug flag, assertions) at the boundaries of

the suspected path.

  1. Rerun the reproduction and read the output.
  2. Fix from evidence, then remove the instrumentation.

4. Prove

  • Rerun the exact reproduction from step 2 and show it now passes.
  • Prefer a regression test that fails without the fix, if tests are allowed.
  • For UI, show a before and after screenshot from the same device.
  • Run the repo's check command (lint and typecheck).
  • Say "passes locally", not "passes".

5. Deliver

  • Commit with the commit skill. Commit only what this task touched. Make one

commit per concern.

  • For a host or app change, deliver only as far as allowed: rebuild the host,

install on a device, or deploy. Then check that the live binary or app runs the new code.

  • Work from a ticket closes with the before/after proof posted on the ticket,

in the format the project's skill defines. Do not change its state.

  • Report:
  • what changed;
  • the proof;
  • anything left behind, recorded with techdebt-mcp if it is debt;
  • anything noticed out of scope.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/luisnquin/nixos-config/dev-flow/badge)](https://verified-skill.com/skills/luisnquin/nixos-config/dev-flow)

Something wrong with this skill or its page? Tell us.