studio
The /studio plugin: repo-native traced development.
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Badge for your README
Markdown
[](/listings/eversluis--studio)HTML
<a href="/listings/eversluis--studio"><img src="/api/listings/eversluis--studio/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
claude plugin marketplace add ~/Projects/studio claude plugin install studio@studio
Then, in any repo, use one of: - `/studio <idea>` (top-down, the preferred mode); - `/studio fix <change>` (bottom-up, small changes only); - `/studio refactor <goal>` (behaviour unchanged, proven by the same tests); - `/studio recode <scope>` (write it fresh from the plan, compare with the old code, and choose per component); - `/studio adopt` (an existing repo); - `/studio audit`.
For Obsidian, see `skills/studio/reference/obsidian.md`.
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
.claude-plugin/marketplace.json · Manifest · 16 lines
- {
- "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
- "name": "studio",
- "description": "The /studio plugin: repo-native traced development.",
- "owner": {
- "name": "Eric Versluis",
- "email": "eric.versluis@paqt.com"
- },
- "plugins": [
- {
- "name": "studio",
- "source": "./",
- "description": "Six-layer traced development as markdown in the repo, readable in Obsidian."
- }
- ]
- }
README.md · README · 50 lines
- # studio
- A Claude Code plugin: `/studio` develops software through six traced layers, kept as concise markdown in the repo's `studio/` folder and read in Obsidian.
- | | EngineeringStudio (app) | /studio (this plugin) |
- |---|---|---|
- | Storage | server and web UI | markdown in the repo |
- | Review | marks, two-step approval | git |
- | Layers | P, R, S, Detailed Design, Analysis, V | P, R, A (Architecture), C (Code), CR (Code Review), V |
- | Link suggestions | embeddings | code facts (paths, imports, per-test coverage) and BM25, judged by the agent |
- | Fits | formal projects, sign-off | everyday development by agents |
- It is the same V-model in a lighter flavour, and it sits next to the app rather than replacing it. The one hard rule is that all code is linked: every public symbol is in a Code entity, and every test is in a Verification entity. The code is never copied: Code and Verification entities embed it live.
- - Process and modes: `skills/studio/SKILL.md`
- - Rules and entity format: `skills/studio/reference/rules.md`
- - Layer agents: `agents/`
- - Helper (stdlib Python 3.11+): `skills/studio/scripts/studio.py`. It reads Python natively and every other language (TS/JS, Go, Java, Kotlin, C#, PHP, Rust, Swift, …) through [Universal Ctags](https://ctags.io) (`brew install universal-ctags`). Coverage runs pytest for Python and any lcov-producing runner for the rest. `studio.py check` runs the repo's own linters and type checkers (ruff, mypy, pyright, eslint, biome, tsc, svelte-check, semgrep: whichever the repo configures) and puts each finding under its Code entity.
- ## Install
- ```bash
- claude plugin marketplace add ~/Projects/studio
- claude plugin install studio@studio
- ```
- Then, in any repo, use one of:
- - `/studio <idea>` (top-down, the preferred mode);
- - `/studio fix <change>` (bottom-up, small changes only);
- - `/studio refactor <goal>` (behaviour unchanged, proven by the same tests);
- - `/studio recode <scope>` (write it fresh from the plan, compare with the old code, and choose per component);
- - `/studio adopt` (an existing repo);
- - `/studio audit`.
- For Obsidian, see `skills/studio/reference/obsidian.md`.
- ## Other models
- The requirements critic, the code mapper and the code reviewer can run on an OpenAI model through the [GitHub Copilot CLI](https://docs.github.com/copilot/how-tos/copilot-cli) (`copilot`), so that a second model reviews Claude's work. Turn it on per repo with `[delegate.<agent>]` in `studio/studio.toml` (see "Delegation to other models" in `skills/studio/SKILL.md`). The CLI uses your `copilot` sign-in, or a token in `COPILOT_GITHUB_TOKEN` (set it in your shell profile, never in a repo). Run `studio.py delegate --list` to see what runs where. When Copilot can't run, the skill falls back to the Claude agent. Delegated runs send repository content to GitHub.
- ## Develop
- ```bash
- python3 -m venv .venv && .venv/bin/pip install pytest pytest-cov
- .venv/bin/python -m pytest -q # the multi-language tests skip without universal-ctags
- ```
- ## Copyright
- Copyright (c) 2026 Eric Versluis. All rights reserved. See `LICENSE`.
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.