SECURITY REVIEW / Addy Osmani
Security and hardening
Starts with a threat model and a three-tier boundary, not with a generic vulnerability list.
“Hardens code against vulnerabilities. Use when auditing an input handler for vulnerabilities, when handling user input, authentication, data storage, or external integrations, or when checking a login flow is safe against the OWASP Top Ten. …”
01 / THE REASONING
Why this made the selection.
- It starts with a threat model: trust boundaries, named assets and a STRIDE pass over each boundary, with abuse cases written next to use cases.
- Its three-tier boundary separates controls that always apply from changes that need human approval (auth logic, CORS, uploads, roles) and from things never to do.
02 / THE REVIEW RECORD
What we actually inspected.
Source review has boundaries.
A clear record is more useful than a “safe” badge.
Material inspected
- skills/security-and-hardening/SKILL.md at 7d65f5e7 (unchanged from 1401c8b8)
- skills/security-and-hardening/references/hardening-patterns.md (changed since 1401c8b8: one zod chain reordered)
- README.md Quick Start and per-skill install note
- LICENSE (MIT)
- README.md (installation command audit at the existing pin)
Our findings
- Its dependency section asks for scripts blocked before first install, the native audit run against the committed lockfile before every release, and never an automatic npm audit fix --force.
- The AI/LLM section treats model output as untrusted input, says the system prompt is not a security boundary, and maps each rule to an OWASP Top 10 for LLM Applications (2025) item.
- Destructive operations on derived paths require the resolved target to sit under an allowlisted root before a delete, move or overwrite.
Not established by this review
- The skill was not executed inside a host agent.
- Not executed: it was not installed or run.
The review applies to the material and revision named here. A newer upstream release can change its behavior.
03 / PUT IT TO WORK
Add a skill to your workflow.
DOCUMENTED COMMAND
npx skills add addyosmani/agent-skills --skill security-and-hardening Run in your terminal to choose your AI tools, project or global scope, and installation method. This short command follows the current upstream version; the review covers the revision shown above.
- Install the named skill.
- Write the threat model before listing controls.
Before you start
- The code that accepts untrusted input
04 / FOLLOW THE EVIDENCE
The source trail.
Our notes are separate from the original resource.
Check upstream before adopting a new version.
- Skill definition ↗Checked
https://github.com/addyosmani/agent-skills/blob/7d65f5e781e819895d0603912290edca6d2cf2f7/skills/security-and-hardening/SKILL.md
Supports: Workflow and stated scope
- Repository licence ↗Checked
https://github.com/addyosmani/agent-skills/blob/7d65f5e781e819895d0603912290edca6d2cf2f7/LICENSE
Supports: Redistribution terms
https://github.com/addyosmani/agent-skills/blob/7d65f5e781e819895d0603912290edca6d2cf2f7/README.md
Supports: install, limitations
https://github.com/addyosmani/agent-skills/blob/7d65f5e781e819895d0603912290edca6d2cf2f7/skills/security-and-hardening/references/hardening-patterns.md
Supports: review, limitations
https://raw.githubusercontent.com/addyosmani/agent-skills/7d65f5e781e819895d0603912290edca6d2cf2f7/skills/security-and-hardening/SKILL.md
Supports: limitations, upstreamDescription, review
https://raw.githubusercontent.com/addyosmani/agent-skills/7d65f5e781e819895d0603912290edca6d2cf2f7/README.md
Supports: review