bdt-status

 ██████╗ ███╗   ███╗███████╗    ███████╗██╗  ██╗██╗██╗     ██╗     ███████╗
 ██╔══██╗████╗ ████║██╔════╝    ██╔════╝██║ ██╔╝██║██║     ██║     ██╔════╝
 ██████╔╝██╔████╔██║███████╗    ███████╗█████╔╝ ██║██║     ██║     ███████╗
 ██╔══██╗██║╚██╔╝██║╚════██║    ╚════██║██╔═██╗ ██║██║     ██║     ╚════██║
 ██████╔╝██║ ╚═╝ ██║███████║    ███████║██║  ██╗██║███████╗███████╗███████║
 ╚═════╝ ╚═╝     ╚═╝╚══════╝    ╚══════╝╚═╝  ╚═╝╚═╝╚══════╝╚══════╝╚══════╝

Reusable AI agent skills for our engineering team — composable instructions that extend coding agents (Claude Code, Cursor, Codex CLI, etc.) with company-specific workflows and best practices.

Skills are a last resort. Anything that can be enforced deterministically — a lint rule, a formatter setting, a pre-commit hook, a CLI check — belongs in linters or tooling, not in a skill. Put it in bmsuisse/devtools (bdt lint, bdt pr, …) so it is checked automatically instead of relying on an agent to remember it. Write a skill only for judgment, conventions and context that tooling cannot express, and link to the matching devtools check where one exists.

---

Available Skills

<!-- SKILLS_TABLE_START -->

Skill	Description
aiochannel-python	Correct patterns for channel-based asyncio concurrency in Python using the aiochannel library. Use this skill wheneve…
autoresearch	Autonomous iterative optimization loop for Python, SQL, ML, and Spark/Databricks. Define a measurable goal; the skill…
azure-cost	Query, analyze, and reduce Azure spending via the Cost Management REST API and Azure CLI. Covers cost breakdowns by s…
bms	Master skill for the bmsuisse platform — routes to the relevant skills based on sub-command. Always enables caveman c…
bms-code-review	BMS code review that fans out one focused subagent per concern (architecture, duplication, security, correctness, age…
bms-frontend-design	The BMS visual identity for internal web apps, built on the shared @bmsuisse/ui (sidebar/nav, form, dialog, and oth…
bms-plan-review	Review an implementation plan before any code is written: checks it against BMS conventions and the existing skills (…
clockify	Query and log time entries in Clockify via its REST API — check a timesheet, look up the workspace/user/project, crea…
code-comments	Write and review code comments so they earn their place instead of adding noise. Use this whenever writing new code (…
codeunit-analyzer	Comprehensive C-AL performance analyzer for Classic Microsoft Dynamics NAV (Navision). Targets Classic NAV anti-patte…
coding-guidelines-python	Apply and enforce Python-specific coding standards. Use alongside coding-guidelines for any Python file — covers typi…
coding-guidelines-sql	SQL and data warehouse coding guidelines for the BME data platform. Use whenever writing, reviewing, or refactoring S…
coding-guidelines-typescript	Apply and enforce TypeScript-specific coding standards. Use alongside coding-guidelines for any TypeScript file — cov…
cross-repo-discovery	Find where a repo lives — locally or in the Azure DevOps org — before assuming a piece of functionality doesn't exist…
data-modeling-dimensional	Dimensional data modeling guide for the Fabricks platform — covers the full pipeline from staging through raw, transf…
databricks-cli	Databricks CLI operations: auth, profiles, data exploration, bundles, and notebook execution. Use this skill for ANY…
databricks-sql-autotuner	Databricks SQL query optimizer and error fixer: analyzes a slow or broken SQL query, rewrites it for speed using SQL-…
design-review	Review a web app or page for visual design quality — layout, spacing, typography, colour/contrast, visual hierarchy,…
deslop	Remove AI slop from code and pull requests. Use this skill whenever the user wants to clean up AI-generated code, rev…
dev-workflow	End-to-end workflow for non-trivial coding tasks: worktree, draft PR early, implement, review, test, publish. Trigger…
duckdb	Guide for working with DuckDB — CLI usage, SQL execution, reading files (CSV, Parquet, JSON, Excel), extensions (http…
ducklake	Guide for working with DuckLake — the open lakehouse format and DuckDB extension that pairs a SQL catalog (metadata,…
fabricks-devops-issues	Orchestrates Azure DevOps "Issue" work-item workflows in the Fabricks.Runtime repo — listing open issues so the user…
fabricks-devops-issues-resolver	End-to-end resolver for Fabricks.Runtime DevOps issues that have already been diagnosed and locally fixed by the fabr…
fabricks-glossary	Use this skill whenever company-specific jargon, acronyms, or domain terminology is needed to answer correctly.
fabricks-sql-analyzer	Analyzes all SQL files in the Fabricks.Runtime repository, builds a dependency DAG, runs performance heuristics, and…
fastapi-azure-auth	Azure Entra ID SSO for FastAPI using cookie-based sessions (MSAL, /login → /callback → session). Trigger on: Azure AD…
fastapi-guideline	Use this skill whenever working with FastAPI — building APIs, adding routes, structuring projects, streaming response…
init-app-stack	Use this skill whenever the user wants to bootstrap, scaffold, or initialize a new full-stack app with a Vite + React…
inter-without-ai-slop	Use Inter as a UI typeface in data-dense B2B apps (dashboards, tables, cockpits) without triggering the "generic AI d…
kendo-ui-angular	Use this skill whenever the user is working with Kendo UI for Angular — including the Data Grid, TreeList, TreeView,…
kendo-ui-react	KendoReact component library patterns and best practices for React developers. Trigger whenever the user works with K…
kendo-ui-vue	Use this skill whenever the user is working with Kendo UI for Vue — including the Data Grid, DropDownList, ComboBox,…
kull-generic-backend	Kull.GenericBackend — an ASP.NET Core middleware (Kull-AG/kull-generic-backend, NuGet package Kull.GenericBackend) th…
maintenance	Periodic repo upkeep for BMS Python (uv) and JS/TS (bun) repos, as one skill with a mode argument: dead-code (find/…
nicegui	Build Python web UIs with NiceGUI — covering layout, widgets, data binding, routing, and the AgGrid data grid. Use th…
playwright-python	Browser automation and visual verification with Playwright in Python. Use this whenever the user wants to drive a bro…
prek	Set up code formatting and pre-commit hooks using prek (fast Rust-based alternative to pre-commit) with prek.toml con…
procrastinate	PostgreSQL-based async task queue for Python using Procrastinate. Use this skill whenever the user is working with Pr…
python-autotuner	Python code optimizer and error fixer: analyzes a Python file or function, rewrites it for speed and quality one chan…
remove-ai-writing	"Detect and surgically fix AI writing patterns in English, German, French, or Italian text. Use this skill whenever t…
rich-data-tables	Use whenever building a UI table, list view, or dashboard panel showing records with amounts, dates, or statuses (inv…
scientific-revision	Use this skill whenever the user wants to verify, revise, or improve a scientific essay, academic paper, or any writt…
spark-connect	Guide for running PySpark code locally against a remote Databricks cluster via Spark Connect (databricks-connect). Us…
sql-optimization	"Universal SQL performance optimization assistant for comprehensive query tuning, indexing strategies, and database p…
tanstack-best-practices	Comprehensive best practices for TanStack libraries in React applications — covering TanStack Query (React Query) dat…
testing-strategy	Decide what kind of test to write and in what order for a backend feature: HTTP-level first, then Playwright e2e for…

<!-- SKILLS_TABLE_END -->

---

Installation

Claude Code Plugin

Add to ~/.claude/settings.json to install all plugins:

{
  "extraKnownMarketplaces": {
    "bmsuisse-skills": {
      "source": {
        "source": "github",
        "repo": "bmsuisse/skills"
      }
    }
  },
  "enabledPlugins": {
    "coding@bmsuisse-skills": true,
    "onetrade@bmsuisse-skills": true,
    "databricks@bmsuisse-skills": true,
    "fabricks-data@bmsuisse-skills": true,
    "writing@bmsuisse-skills": true,
    "ui@bmsuisse-skills": true,
    "caveman@bmsuisse-skills": true,
    "azure-deploy@bmsuisse-skills": true,
    "azure-diagnostics@bmsuisse-skills": true
  }
}

Plugin	Contents
coding@bmsuisse-skills	Python, TypeScript, FastAPI, Postgres, TanStack, and general coding skills
ui@bmsuisse-skills	Data tables, visual design/QA review, and writing style
onetrade@bmsuisse-skills	C-AL / Classic NAV codeunit analysis
databricks@bmsuisse-skills	Databricks CLI + Spark Connect local execution
fabricks-data@bmsuisse-skills	Fabricks / Databricks data skills
writing@bmsuisse-skills	Scientific and academic writing revision
bms@bmsuisse-skills	BMsuisse master skill (/bms, /bms sql, /bms python, /bms data)
caveman@bmsuisse-skills	Ultra-compressed caveman communication mode (~75% token reduction)
azure-deploy@bmsuisse-skills	Azure deployments via azd/terraform/az with validation and RBAC checks
azure-diagnostics@bmsuisse-skills	Azure production diagnostics via AppLens, Monitor, and resource health

Then run these commands in Claude Code to apply the configuration:

# Add the marketplace (one-time setup — alternative to editing settings.json manually)
/plugin marketplace add bmsuisse/skills

# Install the plugins you want (or install all of them, as below)
/plugin install coding@bmsuisse-skills
/plugin install ui@bmsuisse-skills
/plugin install onetrade@bmsuisse-skills
/plugin install databricks@bmsuisse-skills
/plugin install fabricks-data@bmsuisse-skills
/plugin install writing@bmsuisse-skills
/plugin install bms@bmsuisse-skills
/plugin install caveman@bmsuisse-skills
/plugin install azure-deploy@bmsuisse-skills
/plugin install azure-diagnostics@bmsuisse-skills

# After any config change, reload without restarting
/reload-plugins

Command	Description
/plugin	Open the interactive plugin manager (browse, install, enable/disable)
/plugin install <name>@bmsuisse-skills	Install a specific plugin
/plugin uninstall <name>@bmsuisse-skills	Remove a plugin
/plugin enable <name>@bmsuisse-skills	Re-enable a disabled plugin
/plugin disable <name>@bmsuisse-skills	Disable without uninstalling
/reload-plugins	Reload all active plugins without restarting Claude Code

Using the Plugin

Once installed, skills are available as slash commands in Claude Code. Type / to see all available commands, or invoke them directly:

/coding-guidelines-python      # activate Python coding standards
/fastapi-guideline             # load FastAPI best practices
/init-app-stack                # scaffold a Vite+React+TanStack+shadcn / FastAPI+asyncpg app
/databricks-cli                # Databricks CLI operations
/autoresearch                  # start an autonomous research loop
/deslop                        # clean AI slop from code or PRs

Skills also auto-trigger based on context — if a skill's description matches what you're doing, Claude Code may load it automatically without a slash command.

Hooks (automatic tool remapping)

The plugin ships with PreToolUse hooks that automatically rewrite shell commands to use preferred tooling:

You type	Gets remapped to
python script.py	uv run python script.py
python -m pytest	uv run -m pytest
pip install foo	uv add foo
pytest, ruff, ty	uv run pytest, uv run ruff, ...
npm install	bun install
npm install foo	bun add foo
npx foo	bunx foo
yarn add foo	bun add foo
pnpm run dev	bun run dev

This happens transparently — no config needed. The hooks activate for any plugin that includes them (coding, bms, etc.).

Plugin contents

Each plugin bundles a curated set of skills:

Plugin	Skills included	Use case
coding	autoresearch, coding-guidelines-python, coding-guidelines-typescript, deslop, fastapi-guideline, init-app-stack, postgres-best-practices, postgres-test-setup, sql-optimization, tanstack-best-practices	General dev work
ui	rich-data-tables, inter-without-ai-slop, design-review	Data tables, writing style, visual QA
onetrade	codeunit-analyzer	Classic NAV / C-AL analysis
databricks	databricks-cli, spark-connect	Databricks & Spark
fabricks-data	fabricks-glossary, fabricks-sql-analyzer	Fabricks data platform
writing	remove-ai-writing, scientific-revision	Writing & docs
bms	bms (master), coding-guidelines-sql, coding-guidelines-python, coding-guidelines-typescript, sql-optimization, data-modeling-dimensional, fabricks-glossary	All BMSuisse standards via /bms
caveman	caveman, caveman-review, caveman-commit, caveman-help (from JuliusBrussee/caveman)	Token-efficient communication
azure-deploy	azure-deploy (from microsoft/azure-skills)	Azure deployments
azure-diagnostics	azure-diagnostics (from microsoft/azure-skills)	Azure production diagnostics

---

Highly recommended: Caveman mode

We highly recommend Caveman by Julius Brussee. It cuts token usage ~75% by having the agent respond in ultra-compressed style while keeping full technical accuracy. Less noise, faster, cheaper.

bun x skills add https://github.com/JuliusBrussee/caveman

Or via the marketplace:

"caveman@bmsuisse-skills": true

Full credit to Julius for creating and maintaining this — outstanding contribution to the community.

---

skills CLI

The skills CLI reads directly from GitHub — no registration needed.

Install a specific skill

# recommended
bun x skills add https://github.com/bmsuisse/skills --skill <skill-name>
# alternative
npx skills add https://github.com/bmsuisse/skills --skill <skill-name>

Install all skills

# recommended
bun x skills add https://github.com/bmsuisse/skills
# alternative
npx skills add https://github.com/bmsuisse/skills

The CLI auto-detects your agent platform and installs into the right directory:

Platform	Directory
Claude Code	.claude/skills/
Cursor	.cursor/rules/
Codex CLI	~/.codex/skills/
Antigravity	.agents/skills/ or .agent/skills/
Gemini CLI	auto-detected

Browse on skills.sh

https://skills.sh/bmsuisse/skills/<skill-name>

Skills appear on the skills.sh leaderboard automatically based on install telemetry — no manual registration required.

Verify installation

Start a new session and ask your agent something that should trigger the skill.

---

BMS Master Skill

/bms is a master skill that activates the full bmsuisse coding context in one command. It routes to the relevant sub-skills based on a sub-command and always enables caveman communication style.

Sub-commands

Command	Loads	Good for
/bms	coding-guidelines-sql · coding-guidelines-python · coding-guidelines-typescript · fabricks-glossary	General session — all core standards active
/bms sql	coding-guidelines-sql · sql-optimization · fabricks-glossary	SQL authoring, reviews, refactoring
/bms python	coding-guidelines-python	Python development, PR reviews
/bms data	data-modeling-dimensional · fabricks-glossary	Designing or reviewing fact/dim tables

For deep optimization work, chain with the specialized skills:

• SQL benchmarking → /databricks-sql-autotuner
• Python benchmarking → /python-autotuner

How it works

/bms sql
  │
  ├─ reads bms/SKILL.md          (master router, ~2.5k chars)
  │
  ├─ runs load_skills.py sql     (helper script)
  │     └─ concatenates:
  │           coding-guidelines-sql/SKILL.md
  │           sql-optimization/SKILL.md
  │           fabricks-glossary/SKILL.md
  │
  └─ applies all rules + caveman mode → "BMS SQL active."

The helper script at skills/bms/scripts/load_skills.py concatenates the relevant SKILL.md files and prints them as combined content. Each sub-skill remains the single source of truth — no duplication.

Install

Plugin (recommended) — installs bms + all sub-skills as a self-contained plugin:

claude plugin marketplace add bmsuisse/skills
claude plugin install bms@bmsuisse-skills

skills CLI — installs the bms skill only (sub-skills must already be installed):

bun x skills add https://github.com/bmsuisse/skills --skill bms  # recommended
npx  skills add https://github.com/bmsuisse/skills --skill bms  # alternative

Plugin path resolution

When installed as a plugin, load_skills.py resolves sub-skill paths via the CLAUDE_PLUGIN_ROOT environment variable that Claude Code injects at runtime. In the repo, it falls back to the repo-relative path. This means the script works correctly in both contexts without configuration.

---

Adding a Skill

Use skill-creator to guide you through the full lifecycle — drafting, test cases, evaluation, and iteration:

bun x skills add https://github.com/anthropics/skills --skill skill-creator  # recommended
# npx skills add https://github.com/anthropics/skills --skill skill-creator

Then ask your agent: "Help me create a skill for X" and it will walk you through intent capture, writing the SKILL.md, running test cases, reviewing results, and iterating until the skill is ready.

Also helpful is the playwright-cli skill, which you can install like this:

bun install -g @playwright/cli@latest
bunx playwright-cli install --skills

You can give a hint to the agent to use playwright cli to query complex web pages with javascript.

Structure

Each skill lives in its own directory under skills/:

skills/
└── your-skill-name/
    ├── SKILL.md          ← required: instructions + YAML frontmatter
    ├── scripts/          ← optional: helper scripts
    ├── references/       ← optional: reference documentation
    └── assets/           ← optional: templates, icons, etc.

SKILL.md format

---
name: your-skill-name
description: What the skill does and when to trigger it. Be specific — this is how the agent decides whether to use it.
---

# Your Skill

Instructions for the agent to follow...

Workflow

1. Create a branch: git checkout -b skill/<skill-name>

2.