ryanzillini

AI Engineering Skills — Security skill for Claude Code

Security community

Focused Claude Code / Cursor skills for systematic AI-native engineering: codebase discovery, security review, planning, and validation.

How to install AI Engineering Skills

This entry records only its repository, not the path inside it, so there is no exact command to give. Open ryanzillini/ai-engineering-skills and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What AI Engineering Skills does

Focused Claude Code / Cursor skills for systematic AI-native engineering: codebase discovery, security review, planning, and validation. Durable artifacts, progressive disclosure, judgment over generation.

Alternatives in Security

Pal — Multi-model AI integration — chat, debugging, code review, planning, security audit 11.3k ★
Harden — Run the harden skill to perform a systematic project hardening audit 1k ★
Maestro Knowledge — Intent-driven knowledge-store and Run knowledge lifecycle management — audit/prune, stage candidates (with sig 530 ★

README

AI Engineering Skills

Focused, portable skills for systematic AI-native engineering with Claude Code, Cursor, and Codex.

These are not generic prompts. Each skill is a focused procedure that produces a durable, checkable artifact. The model proposes; the skill and the human decide what is accepted.

Philosophy

One focused skill per scenario. Composability over kitchen-sink prompts.
Progressive disclosure: short SKILL.md, heavier detail in references/ when needed.
Durable outputs (markdown corpus, threat model, plan, validation report) so the next session starts oriented instead of re-scanning.
Evidence over assertion. Tag claims as extracted, inferred, or ambiguous.
Defaults are often too slow or unreliable for production judgment. These skills encode the review, discovery, and fail-closed patterns that matter.

Skills

Skill	When to use	What it produces
`codebase-discovery`	Onboarding to an unfamiliar codebase, or building a reference before deep work	Structured corpus: architecture, data models, patterns, critical files
`security-review`	Reviewing changes or a module for realistic failure modes	Prioritized findings with trust boundaries and evidence
`planning`	Turning an ambiguous requirement into a constrained plan	Plan with verification steps and second-order checks
`validation`	Checking generated or proposed work against contracts or golden sets	Validation report or eval harness outline

Usage

Place the skills in a location your agent reads:

Claude Code: .claude/skills/ or ~/.claude/skills/
Cursor: .cursor/skills/, .agents/skills/, or the Claude/Codex compatibility paths
Codex: .agents/skills/

Invoke explicitly with `/skill-name` or let the agent match on the description.

These skills work best when the agent has tools to read the filesystem and run limited commands. They do not replace human judgment on architecture, security, or what