One-click install
npx skills add https://github.com/nodatall/primedirective --skill repo-sweep
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: repo-sweep
Source: https://github.com/nodatall/primedirective/tree/main/skills/repo-sweep
Command: npx skills add https://github.com/nodatall/primedirective --skill repo-sweep

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Repositories that pass local tests and standard code reviews often still hide critical production risks, from misconfigured auth and exposed admin endpoints to architectural drift and unhandled edge cases. This skill eliminates that blind spot by running a comprehensive, first-principles no-edit audit of your entire codebase before any fixes are applied, ensuring no hidden risks are missed and no changes are made without your explicit approval.

Core Features & Use Cases

  • First-Principles No-Edit Pre-Pass: Starts with a mechanism-level understanding of your repo to detect production risks even when the code "works" locally, carrying a concise audit thesis through the entire review process.
  • Comprehensive Mandatory Review Chain: Forces full coverage of security, configuration, API surface, go-live readiness, and dependency checks, rather than shortened provider-specific subsets, with optional modifiers for Pro analysis, swarm maintainability reviews, and dependency supply-chain scans.
  • Structured Reporting with Approval Gates: Emits a clear, severity-ordered repo-wide report of verified findings before any fixes are applied, and only proceeds with repairs after you give explicit approval, with support for bounded repair/resweep loops via goal mode.
  • Use Case: Ideal for engineering teams preparing a service for production deployment, conducting a full repo health check before a major release, or investigating unexplained production issues in a mature codebase.

Quick Start

Invoke the repo-sweep skill to run a full first-principles audit of your current repository and receive a structured risk report with approval required before any fixes are applied.

Frequently Asked Questions about repo-sweep

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

FAQPage Schema
How do I run a code review to catch production risks that local testing misses?

A comprehensive repo audit applies a first-principles no-edit review to catch production risks like misconfigured auth and exposed endpoints that pass local tests. It forces full coverage of security, configuration, API surface, and dependency supply-chain checks before deployment.

When do I need a dependency audit and supply-chain scan before production deployment?

You need a dependency audit and supply-chain scan when preparing a service for production deployment or a major release. It verifies go-live readiness by checking for vulnerabilities in your dependency tree that standard code reviews often overlook.

Can I review my codebase for maintainability without applying any automatic fixes?

Yes, a no-edit pre-pass audit reviews your codebase for maintainability and production risks without applying any automatic fixes. It emits a severity-ordered report of verified findings and requires your explicit approval before proceeding with any code modifications.

What is the best way to verify go-live readiness for a mature codebase?

The best way to verify go-live readiness is a full repo health check using a first-principles audit. This process enforces a mandatory review chain covering security, configuration, API surface, and dependencies to ensure no hidden risks remain before deployment.

Does the repo audit process support bounded repair loops after identifying issues?

Yes, the audit process supports bounded repair and resweep loops via goal mode after you explicitly approve the proposed fixes. This ensures you maintain control over code modifications while systematically resolving verified findings from the initial report.

What are the limitations of standard code reviews for detecting unexplained production issues?

Standard code reviews often miss architectural drift, unhandled edge cases, and misconfigured endpoints because they lack a mechanism-level understanding of the entire repo. A first-principles audit is needed to carry a concise thesis through the full codebase review.