What it does
This entry is a single SKILL.md file from the Anthropic-Cybersecurity-Skills library published by mukul975. The file's subject is detecting broken object property-level authorization. The library holds 818 structured cybersecurity skills across 34 security domains, each following the agentskills.io open standard, and each skill is mapped to the frameworks relevant to its type, such as MITRE ATT&CK and NIST CSF 2.0.
It suits security reviews that examine authorization at the level of individual object properties, where an AI agent is asked to guide that work. The source does not describe the skill's steps in more detail.
The skill is used by cloning the repository and pointing an agent at it. The library states that it works with Claude Code, GitHub Copilot, OpenAI Codex CLI, Cursor, Gemini CLI and any agentskills.io-compatible platform.
Limitations and requirements
The library includes offensive and dual-use techniques and is intended only for authorized penetration testing, security research, defense and education. Use it only against systems you own or have explicit written permission to test, and comply with applicable laws and rules of engagement. The publisher states that the project is community-created, is not affiliated with Anthropic PBC, and that the user is solely responsible for how the skills are used.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
Sources
- AgentSkillsDir first seen 13 Nov 2026 · last seen 18 Nov 2026 · its page on the source
Review source
As shown on agentskillsdir.com when we last read it