Contribute upstream
Corrections and new skills belong in the canonical community repository, so every downstream catalog can benefit.
Contribution guideSecurity Skills Atlas turns a community-maintained repository into a navigable reference for practitioners and AI agents—without hiding the source, executing its helpers, or overstating its provenance.
The source library and this catalog are not affiliated with, sponsored by, or endorsed by Anthropic. “Anthropic” appears in the upstream repository name only. The catalog uses neutral branding to avoid implying an official relationship.
817 canonical skills across 46 current security domains. Each record keeps its procedures, metadata, framework mappings, references, scripts, and assets tied to its source folder.
Each skill is a portable folder led by SKILL.md: YAML metadata plus Markdown instructions and optional supporting resources. The structure follows the open conventions described by agentskills.io.
Helper scripts are treated as untrusted source text. The data build reads and hashes them; the website syntax-highlights them for review. It never imports, evaluates, launches, or schedules their contents.
An agent installs the library, selects a relevant skill, reads its prerequisites and workflow, and applies the guidance within the permissions of its own environment. A skill is instruction, not automatic authorization.
Run the upstream library’s standard install command in a supported agent-skills client.
npx skills add mukul975/Anthropic-Cybersecurity-SkillsCorrections and new skills belong in the canonical community repository, so every downstream catalog can benefit.
Contribution guideThe upstream repository and its canonical skills declare the Apache License 2.0. Attribution and license notices remain available with the source.
Read the licenseSearch by mission, technology, tag, or control framework.
Open the skills catalog