PlugWiki

principal-engineering skill (riekelt): the core rules for engineering with Claude

Core rules for coding agents: check the real code first, fail loudly, and never claim done without proof.

Installs
20.4K
on skills.sh
Stars
7
on GitHub
Updated
Sep 2026
last repo commit

principal-engineering is the foundation skill of the principal-engineer set. It makes Claude fill in a four-part checkpoint (Grounded, Blast radius, Invariants, Verify) from the real repository before a non-trivial change, and it carries six hard rules such as no silent error swallows and never printing secret values. Its description says to use it for any non-trivial engineering work, even a routine-looking one. The other ten skills in the set each name it as required background.

Good for

  • Developers who want Claude to check the real code before it changes anything
  • Teams that want a written checkpoint before risky changes
  • Anyone installing the other ten principal-engineer skills, which all build on this one

Install principal-engineering

Claude Code

  1. 1

    Copy this prompt.

    Install the principal-engineering skill in Claude Code. Run this command:
    
    /plugin marketplace add riekelt/principal-engineer
    /plugin install principal-engineer@principal-engineer
    
    Then check that it loaded and tell me how to use it.
  2. 2

    Paste it into Claude Code. It shows the change and asks before running it.

Pick one route. The plugin installs all 11 skills in the repo, plus one agent (change-verifier) and two commands (/verify and /preflight), counted from the git tree. The skills CLI and folder routes below install only this skill; the other ten each name it as required background. There is no Claude app upload route: the README documents only the Claude Code plugin and the other-agents routes.

From the project README. Source.

What's in the skill

When it loads

Use when doing any non-trivial engineering work - implementing, debugging, refactoring, configuring, operating, or investigating why a system misbehaves - or any change where being wrong has a cost. Encodes the evidence-over-theory discipline, the hard safety rules, and the pre-change checkpoint. Use whenever code, data, or infrastructure is about to change or must be understood before it can, even if the task looks routine or is only "find out why". Foundation for the sibling skills.

Sections in SKILL.md

  1. Overview
  2. When to invoke
  3. Scope limits
  4. Mandatory checkpoint before a non-trivial change
  5. Hard rules
  6. Risk tiers set the rigor
  7. The rule lifecycle
  8. Common mistakes

Key facts

Installs on skills.sh
20,370 (Nov 2, 2026)
Last repo commit
Sep 11, 2026

What it includes

  • Fills Grounded, Blast radius, Invariants and Verify before the first line

    Before a non-trivial change Claude writes four fields in its notes. Each one has to come from the code and data, not from memory.

    Read · Pre-change checkpoint

  • Loud failures, immutable migrations, named proof, no secret reads, guarded deletes

    The rules hold in every repository: no silent error swallows, applied migrations are never edited, a done claim names what was checked, secret values are never read or printed, destructive operations get a look first, and evidence beats theory.

    Read · Six hard rules

  • The project declares its top-tier paths and the tier sets how much proof

    Money paths, stored user data or a safety gate can be declared top tier. When the project does not say, Claude is told to ask what the system must never get wrong.

    Read · Risk tiers

  • A lesson that bites twice becomes a written rule with its provenance

    Something that bites once is learning. Twice, it becomes a rule with what happened, when and how to avoid it. A rule whose cause is fixed gets retired.

    Read · Rule lifecycle

  • Names which of the ten sibling skills to load for each kind of task

    Ten rows, from starting a change (grounding-before-coding) to adding a package (adding-dependencies).

    Read · Routing table

  • Not a style guide; the repository rules and CLAUDE.md outrank it

    Formatting, naming taste and framework choice belong to the repository. It does not decide what to build.

    Read · Scope limits

From the official docs

Permissions

Low risk

principal-engineering is instructions for Claude and the plugin route adds an agent and two commands. Our advice: review the diff after each session and keep your tool's permission prompts on.

What it can do

  • Read your data
  • Create and change things
  • Delete things
  • Read-only mode

Watch out for

  • The plugin route adds an agent that can run shell commands

    change-verifier lists the tools Read, Glob, Grep and Bash, and the plugin adds the /verify and /preflight commands. The repo tree has no hooks or MCP config files. Use the skills CLI route to install only this skill. github.com

Technical details
Turning on read-onlySKILL.md describes no read-only mode.
Sign-in and keysSKILL.md names no account, key or token.
Scopes
  • Claude reads your repository and writes a checkpoint in its notes before a non-trivial change

  • The plugin route also installs 10 other skills, one agent and two commands

Questions

What does principal-engineering do?

It makes Claude ground a non-trivial change in the real code, fill a four-field checkpoint, follow six hard rules and name the proof before saying done. It is the base the other ten principal-engineer skills build on.

Do I need it if I install another principal-engineer skill?

Yes. Each of the ten other skills opens with "REQUIRED BACKGROUND: the principal-engineering skill". The skills CLI and folder routes in the install box add it for you, and the plugin route installs all eleven skills.

Is it a style guide?

No. SKILL.md says formatting, naming taste and framework choice belong to the repository, and that CLAUDE.md and repository rules outrank everything in the skill.

Does it need an API key or an account?

SKILL.md names no key, account or script. The skill folder is one file, SKILL.md.

Does it work in the Claude app or Codex?

The README documents the Claude Code plugin commands and, for other agents, pointing the platform's plugin loader at plugins/principal-engineer/ or symlinking the skill folders. It documents no Claude app upload route.

Is it safe to install?

The skill folder bundles no scripts. It tells Claude to look before deleting, ask before restarting live services and never print secret values. The plugin route also adds an agent that can run shell commands, covered under "Watch out for".

Information

updated Nov 2, 2026

Folder
plugins/principal-engineer/skills/principal-engineering

riekelt/principal-engineerskills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 4
  1. principal-engineering SKILL.md · github.com
  2. README · github.com
  3. plugin.json · github.com
  4. change-verifier agent · github.com

Updated Nov 2, 2026.