#安全

3 tutorials, projects and news items tagged 安全.

Find, verify and report bugs within authorized scopePick

Use this when you want to run a security assessment within scope you own or have written authorization for (reviewing your own project, an authorized penetration test, a CTF challenge, or hardening against the OWASP checklist), and want your AI agent to follow a disciplined, reproducible staged process that produces evidence and a report. Based on the open-source Claude-BugHunter (MIT / CC BY 4.0): confirm the authorized scope first, then map the exposed surface, check it against a checklist, verify each finding one by one, and write the report in the platform's format. Test only authorized targets. No post-exploitation or attack-infrastructure development. Use case: Within authorized scope, have your AI agent hunt for vulnerabilities, verify them, and write the report using a disciplined workflow.

Skill·Intermediate·4 minskill安全漏洞渗透测试
Have your AI agent check your site's hidden states, it may expose a serious flaw

How orders and accounts change state is often scattered across many endpoints, and bugs hide in the changes nobody expected. Includes a prompt that has your AI agent read the code only and list suspicious state changes.

Experience·Beginner·1 min·Members安全漏洞状态机代码审计
What Claude Code's Auto Mode Actually Blocks for You

Auto Mode has a classifier model check each action before it runs. By default it blocks risky steps like running downloaded scripts right away, production deploys and force pushes. If it keeps getting blocked, it asks you again. Here is how it compares with turning all permissions off.

News·Beginner·3 min·Nov 15, 2026claude codeauto modepermissions权限