Auth0's official agent skill: a router that detects your framework and intent, then loads the right Auth0 guidance for login, MFA, Organizations, tenant audits, debugging or provider migration.
Security and Hardening Skill
Addy Osmani's skill for security-first web development: threat model first, always/ask-first/never rules, hardening controls from injection and SSRF to supply chain, privacy and LLM features, and a review checklist.
Install
Other ways to install
About
- Threat model first, then a three-tier always / ask-first / never boundary system
- Hardening controls for injection, XSS, sessions, headers, CORS, uploads and SSRF
- Dependency and supply-chain checks with the native package-manager audit
- Guidance for personal data (GDPR, CCPA) and AI or LLM features
- Review checklist, red flags and verification steps
Security and Hardening is part of Addy Osmani's addyosmani/agent-skills collection of engineering skills. It treats security as a constraint on every line that touches user data, authentication or external systems rather than a phase at the end, and tells the agent to apply it whenever a feature accepts untrusted input, manages sessions, calls third-party services, adds dependencies or handles personal data under GDPR or CCPA.
Its process starts with a threat model, then applies a three-tier boundary system. Always: validate input at the boundary, parameterise queries, encode output, use HTTPS, hash passwords with bcrypt, scrypt or argon2, set security headers, use httpOnly, secure, sameSite cookies, and run the package manager's own audit against the lockfile before release. Ask first: new auth flows, new categories of sensitive data, new integrations, CORS changes, upload handlers, rate-limit changes and elevated roles. Never: commit secrets, log sensitive data, trust client-side validation, use eval() or innerHTML with user data, or keep auth tokens in localStorage.
The hardening section then works through injection, XSS and access control, sessions, headers and CORS, uploads, server-side request forgery, destructive operations on derived paths, rate limiting, secrets, dependencies and supply chain, privacy, and AI or LLM features. It ends with a review checklist, common rationalisations to reject, red flags and verification steps. MIT licensed.
SKILL.md
The instructions your agent loads, as published by Addy Osmani.
Related Skills
See all Security and Hardening alternativesBetter Auth's official skill for hardening a Better Auth deployment: secret management, rate limiting, CSRF protection, trusted origins, session and cookie security, OAuth token encryption, IP tracking and audit logging.
Skill: Better Auth: Two-Factor Auth
by Better Auth
Adds real 2FA to a Better Auth app — TOTP, email/SMS OTP, backup codes, trusted devices — including the sign-in interrupt and the encryption-at-rest details.
Skill: Better Auth: Email & Password
by Better Auth
Credential auth done to spec in Better Auth: verification email, reset flow, token and session handling, password policy and custom hashing.