repo-rules

Enforces project-wide editing rules for the dlthub-start repository before file modifications.

1|Updated May 13, 2026
One-click install
npx skills add https://github.com/dlt-hub/dlthub-start --skill repo-rules-dlt-hub
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: repo-rules
Source: https://github.com/dlt-hub/dlthub-start/tree/main/.claude/skills/repo-rules
Command: npx skills add https://github.com/dlt-hub/dlthub-start --skill repo-rules-dlt-hub

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Contributors to the dlthub-start repository can accidentally hand-edit auto-generated AI workbench files, causing make check-ai failures and lost changes on the next regeneration. This Skill ensures those rules are read before any file in the codebase is modified. ## Core Features & Use Cases - Mandatory pre-edit rules: Instructs the agent to always read project-wide rules before modifying any file in the repo. - Generated-file protection: Flags the vendored AI workbench files under src/create_dlthub_workspace/scaffolds/minimal_workspace/_agents/** (Claude, Cursor, Codex) as auto-generated and off-limits to direct edits. - Correct change workflow: Directs contributors to edit the upstream dlt-hub/dlthub-ai-workbench source and regenerate via make update-ai or make generate-ai against the pinned WORKBENCH_REF. - Use Case: Before refactoring a scaffold template, the agent checks this Skill, avoids editing the vendored files directly, and instead regenerates them from the pinned workbench ref. ## Quick Start Read the repo-rules skill before modifying any file in the dlthub-start repository to confirm whether the target file is auto-generated.

Frequently Asked Questions about repo-rules

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I modify the AI workbench files in dlthub-start?▼

Do not edit the vendored files under src/create_dlthub_workspace/scaffolds/minimal_workspace/_agents directly. Edit the upstream dlt-hub/dlthub-ai-workbench source, then regenerate with make update-ai or make generate-ai and commit the regenerated scaffold.

Why does make check-ai fail after editing scaffold files?▼

The check fails because the vendored AI workbench files are auto-generated from the pinned WORKBENCH_REF, so any hand edit creates drift. Regenerate the files with make generate-ai or make update-ai to restore consistency.

Which files in dlthub-start are auto-generated?▼

The AI workbench files under src/create_dlthub_workspace/scaffolds/minimal_workspace/_agents/** are auto-generated, covering the Claude, Cursor, and Codex agent configurations. They are produced from the pinned WORKBENCH_REF of the upstream workbench repo.

When should I run make update-ai versus make generate-ai?▼

Use make update-ai after changing the upstream dlt-hub/dlthub-ai-workbench source to pull and regenerate content. Use make generate-ai to regenerate against the already-pinned WORKBENCH_REF without bumping it.