How does Repo Guardrails prevent database leakage?

The skill enforces a rule that only the apps/api directory can access the database, blocking any direct database calls from other layers like apps/bot.

Can I still run git commands with this skill active?

Read-only commands like git status and git diff are allowed to run automatically, but any destructive or external commands require your prior approval.

Does this skill prevent the AI from adding new dependencies?

Yes, it specifically blocks the addition of dependencies that have not been explicitly requested or authorized in the current task scope.

What happens if the AI tries to change too many lines of code?

If a projected change exceeds approximately 200 lines, the skill triggers a guardrail that pauses the operation and requests manual confirmation to ensure safety.

Repo Guardrails

Name: Repo Guardrails
Author: Hakuya5247

byHakuya5247

Collaboration Tools

Enforces strict architectural boundaries and development scope limits to maintain repository integrity.

About

Repo Guardrails is a governance-focused skill designed to keep AI-driven development within the bounds of your project's architecture. It prevents unauthorized database access by restricting interactions to specific API layers, mandates a contract-first approach by anchoring API definitions to shared packages, and strictly controls the scope of changes to prevent unintended side effects. By monitoring diff sizes and command execution, this skill ensures that Claude stays focused on assigned tasks while adhering to the technical standards of your monorepo.

Key Features

Diff size monitoring to pause operations if changes exceed 200 lines
Scope control to prevent unauthorized file and dependency changes
Strict enforcement of database access through designated API packages
Safe command execution logic with manual approval for destructive actions
Contract-first validation for API input and output definitions
0 GitHub stars

Use Cases

Ensuring compliance with shared API contracts across multiple services
Maintaining architectural separation between API and Bot layers in a monorepo
Preventing 'scope creep' during automated code refactoring sessions

About

Key Features

Diff size monitoring to pause operations if changes exceed 200 lines
Scope control to prevent unauthorized file and dependency changes
Strict enforcement of database access through designated API packages
Safe command execution logic with manual approval for destructive actions
Contract-first validation for API input and output definitions
0 GitHub stars

Use Cases

Ensuring compliance with shared API contracts across multiple services
Maintaining architectural separation between API and Bot layers in a monorepo
Preventing 'scope creep' during automated code refactoring sessions