Skip to content
Back to skills

Nexgen Discipline

ASecurity

Use for ANY code change, before writing code. A search-first, smallest-diff, evidence-based workflow. The owner is not a programmer, so do not rely on them to catch mistakes.

  • 2 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 18, 2026
ai-agentspythongoshellcode-review

Security analysis

A100/100

Scanned October 18, 2026

npx -y skills add matteopasseri407/NeXgen-Engine --skill nexgen-discipline --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Nexgen Discipline?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Nexgen Discipline
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/matteopasseri407-nexgen-discipline/badge)](https://www.skillsdirectory.com/skills/matteopasseri407-nexgen-discipline)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: nexgen-discipline
description: Use for ANY code change, before writing code. A search-first, smallest-diff, evidence-based workflow. The owner is not a programmer, so do not rely on them to catch mistakes.
---

# NeXgen discipline

The rules are in `03-INFRA/engineering-rules.md`. This is the order to apply them in.

0. **New project?** Run `python 03-INFRA/scripts/discipline.py init --smoke "<command>"` once, then `report` shows the owner the state.
1. **Interview if unclear.** When the request is vague or a design choice is open, grill the owner before building: numbered
   rounds, every question with your recommended answer worded so that "yes" accepts it, facts looked up by you and never asked, and
   no work until the owner confirms. Otherwise **restate** in two lines what you will do and what you will not touch.
2. **Search** (`rg`) for existing code that does it; list what you will reuse.
3. **Propose** the minimal diff. If it needs a dependency, script language, data-format change or CI/hook
   change, say so in plain words and wait for the owner's yes (unattended: write the proposal down and do not do it).
4. **Test first** for a bug: build a command that goes red on exactly this bug before touching the code, and commit it with the fix as `fix: ...`
   (the gate checks that the test fails on the old code). A behavior test for a feature. Critical code never ships without a test.
5. **Implement** small. Logic in Python, shell only as a thin launcher.
6. **Check:** `python 03-INFRA/scripts/discipline.py check` and the project's tests. Show the output.
7. **Score:** `python 03-INFRA/scripts/discipline.py score` is switched off and says so: no change needs a review and no gate checks a receipt. A review is the owner's call, not a step: if they ask for one, `review` writes the brief, `council code-review` runs it, and `approve --reason` records the owner's own decision, which you never run on your own.
8. **Self-critique:** what could break, what could you delete, what did you not verify.
9. **Report** in the owner's language, plain words: what changed, files, commands run, risks, next step.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.