Review And Architecture Skills

This repository contains four skills as sibling directories:

• diff-review
• repo-audit
• architecture-standards
• fallow

Contents

diff-review

• Reviews local git diffs for bugs, security issues, and code quality before pushing.
• Includes the skill definition, agent config, stack references, review process references, static-analysis guidance, and scripts/review-preflight.sh.

repo-audit

• Runs full repository audits across correctness, security, architecture, performance, and maintainability.
• Includes the skill definition, agent config, stack references, audit process references, static-analysis guidance, and scripts/audit-preflight.sh.

architecture-standards

• Provides architecture guidance for design, refactoring, scaffolding, and architectural code review.
• Includes the skill definition, agent config, architecture reference packs for current-state diagnosis, target-state design, refactor design, static-analyzer policy, and scripts/architecture-preflight.sh.

fallow

• Guides free-version Fallow adoption, configuration, reruns, remediation, and interpretation for TypeScript/JavaScript codebase intelligence.
• Includes the skill definition, agent config, and Fallow workflow, analysis primitive, package internals, and quality benchmark references.

How They Work Together

• diff-review and repo-audit are the investigation skills. They focus on finding bugs, risks, regressions, and weak design decisions in real code.
• architecture-standards is the design and boundary guide. It helps shape remediation, evaluate architectural quality, and keep fixes proportionate to the actual problem.
• fallow is the free-version static-analysis signal source for TypeScript/JavaScript repos. It helps surface dead code, duplication, health hotspots, audit gates, and baseline policy without relying on trial, paid, licensed, hosted, pro, or runtime-coverage features.
• If architecture-standards is installed alongside the review skills, it can be used to strengthen investigation quality, architectural reasoning, and fix recommendations.
• If fallow is installed alongside repo-audit, Fallow output can be recorded as audit evidence instead of being presented as raw tool output alone.
• The skills still work independently if companion skills are not installed.

Layout

.
├── architecture-standards/
│   ├── README.md
│   ├── SKILL.md
│   ├── agents/openai.yaml
│   ├── references/
│   │   ├── architecture-shapes.md
│   │   ├── current-state-diagnosis.md
│   │   ├── layer-standards.md
│   │   ├── target-state-design.md
│   │   └── ...
│   └── scripts/
│       └── architecture-preflight.sh
├── diff-review/
│   ├── README.md
│   ├── SKILL.md
│   ├── agents/openai.yaml
│   ├── references/
│   │   ├── nextjs.md
│   │   ├── review-workflow.md
│   │   ├── static-analysis.md
│   │   └── ...
│   └── scripts/
│       └── review-preflight.sh
├── fallow/
│   ├── README.md
│   ├── SKILL.md
│   ├── agents/openai.yaml
│   └── references/
│       ├── analysis-primitives.md
│       ├── fallow-workflows.md
│       └── ...
├── repo-audit/
│   ├── README.md
│   ├── SKILL.md
│   ├── agents/openai.yaml
│   ├── references/
│   │   ├── nextjs.md
│   │   ├── audit-workflow.md
│   │   ├── static-analysis.md
│   │   └── ...
│   └── scripts/
│       └── audit-preflight.sh
└── README.md