No red flags found in any of the 11 categories — no credential harvesting, no data exfiltration, no curl-pipe-shell installer. Scanned against the SlowMist agent-security taxonomy, refreshed every 8 hours. Full audit →
by XSafeAI · Codex Skill · ★ 160
Last updated: · Indexed by AgentSkillsHub · Auto-synced every 8h
🔒 Is XSafeClaw safe to install? View the security audit →
English · 中文文档 XSafeClaw: Monitor and Secure Your Agents 🚀 🔥🔥 Now supporting OpenClaw, Hermes, and Nanobot! 🔥🔥 AI agents are not just new software. They are software that can be talked into doing dangerous things. As agents move from chatbots to active systems that browse the web, execute code, and operate inside real workflows, we have handed language models the keys to our infrastructure before figuring out how to keep them on the rails. This breaks traditional security assumptions entirely. In conventional systems, behavior is defined in code. In agents, behavior emerges at runtime from instructions, retrieved content, memory, and long decision loops. An attacker no longer needs to exploit a bug. They can manipulate the agent's reasoning, redirect its trajectory, or turn small permissions into larger ones over time. Prompt injection
| Stars | 160 |
| Forks | 8 |
| Language | Python |
| Category | Codex Skill |
| License | MIT |
| Quality Score | 63.5439690760299/100 |
| Open Issues | 1 |
| Last Updated | 2026-07-10 |
| Created | 2026-03-10 |
| Platforms | python |
| Est. Tokens | ~17k |
Looking for a XSafeClaw alternative? If you're comparing XSafeClaw with other codex skill tools, these 6 projects are the closest alternatives on Agent Skills Hub — ranked by topic overlap, star count, and community traction.
Deterministic safety solutions for probabilistic AI agents
A comprehensive reference for securing Large Language Models (LLMs). Covers OWASP GenAI Top-10 risks, prompt i
Curated, security-first OpenClaw skills (Markdown-based). Security audit skills - detect prompt injection, sup
Security toolkit for AI agents. Scan your machine for dangerous skills and MCP configs, monitor for supply cha
Security toolkit for AI agents. Scan your machine for dangerous skills and MCP configs, monitor for supply cha
🛡️ A curated list of resources on agent skills security: attacks, defenses, frameworks, and benchmarks for se
Explore other popular codex skill tools:
XSafeClaw is Introducing XSafeClaw: The Open-Source Agent Safety Platform from Fudan University. It is categorized as a Codex Skill with 160 GitHub stars.
XSafeClaw is primarily written in Python. It covers topics such as agent-safety, agentic-ai, ai-safety.
You can find installation instructions and usage details in the XSafeClaw GitHub repository at github.com/XSafeAI/XSafeClaw. The project has 160 stars and 8 forks, indicating an active community.
XSafeClaw is released under the MIT license, making it free to use and modify according to the license terms.
The top alternatives to XSafeClaw on Agent Skills Hub include Sponsio, LLMSecurityGuide, openclaw-skills-security. Each offers a different approach to the same problem space — compare them side-by-side by stars, quality score, and community activity.
Grades come from a rule-based scan built on the SlowMist agent-security taxonomy, covering 11 red-flag categories including credential harvesting, data exfiltration, and curl | sh installers. It is a first-layer scan, not a manual audit — we say so rather than overstate it.
The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.
Sources & who's responsible: