Intelligence Digest
github安全推送
GitHub 安全开源项目与漏洞监控情报推送
来自 GitHub Issues、仓库检索和关键词命中的 CVE / RCE / POC 动态。
akash-osmsec/CVE-2026-44262
language: Blade
manoelprovider20-lgtm/merge-authz-test
authz test (CVE-2026-1999 siblings)
zeroc00I/CVE-2026-63030
That wasnt tested on a vulnerable target yet
codeb0ssx/CVE-2026-15583-PoC
Academic proof-of-concept demonstrating CVE-2026-15583 for authorized security research. | topics: codeb0ss, cve-2026-15583, cve-2026-15583-poc, grafana, realcodeb0ss, thecodeb0ss | language: Python
codeb0ssx/CVE-2026-46442-PoC
Academic proof-of-concept demonstrating CVE-2026-46442 for authorized security research. | topics: codeb0ss, cve-2026-46442, cve-2026-46442-poc, flowise-ai, realcodeb0ss, thecodeb0ss | language: Python | homepage: https:...
CerberusMrXi/WordPress-KeepInMind-CVE-2026-9271-Exploit
WordPress KeepInMind CVE-2026-9271 Exploit - Tool detecting stored XSS in KeepInMind plugin v0.8.4.2 and below. Built by Sudeepa Wanigarathna, it simulates CSS injection to hijack admin accounts. Features safe testing, a...
MiaPatsune/cve-2026-43499
language: HTML
bekwiner/cve-2026-47777
oscerd/CVE-2026-48204
Reproducer for CVE-2026-48204: Apache Camel camel-mongodb-gridfs gridfs.* header injection overriding the GridFS operation (enumerate/read/delete files) from an unauthenticated HTTP request (fixed in 4.14.8/4.18.3/4.21.0...
Industri4l-H3ll-Xpl0it3rs/CVE-2026-33017-Langflow-RCE
CVE-2026-33017 Exploit | by infrar3d | topics: exploit | language: Python
优先展示中文安全团队维护的开源项目，兼顾工具落地场景和最近更新时间。
jackbreak66/js-audit-skills
面向红队的 AI Agent 驱动前端 JavaScript 全链路安全审计 Skills | topics: ai-agent, javascript, penetration-testing, redteam, security-audit | language: Python | stars: 2 | forks: 0 | updated 2026-06-12T01:08:35Z | pushed 2026-06-08T04:21...
scf131024-Night-Star/-Counter-Consensus-Lens
反共识透镜（Counter-Consensus Lens） 是一个 TRAE SOLO Skill，输入任意可辩论话题，即可生成一份结构化的批判性认知简报，包含：主流共识地形、高质量反对观点、14维度认知盲区分析、信源上下文增强。 适合需要快速建立全面认知框架的用户——辩论准备、学术综述、决策红队推演、商业调研、内容创作。 | stars: 1 | forks: 1 | updated 2026-06-16T10:29:52Z | pu...
Xiaohei7X7/CompeteInsight
CompeteInsight 是一个面向竞品研究的 AI 工作台。它把一个模糊的调研问题，拆解成可追踪的搜索、抓取、证据整理、红队审查和报告生成流程，最后输出结构化证据、竞品矩阵、Markdown 报告，以及可继续追问的 AI 分析助手。 | language: Python | stars: 3 | forks: 0 | updated 2026-07-16T15:21:53Z | pushed 2026-06-16T04:52:00...
benben951/aas-attack
AI Agent Security (Kaggle): 逆向评分系统 + 证明本地靶子结构无解 + 自适应红队搜索 | stars: 0 | forks: 0 | 2026-06-16T16:13:24Z
CN-big-cabbage/skill-promptfoo
LLM 提示词测试、模型评估与红队安全扫描框架，通过声明式 YAML 配置驱动自动化评估流水线，支持 CI/CD 集成，被 OpenAI 和 Anthropic 内部使用 | stars: 0 | forks: 0 | updated 2026-06-17T03:08:19Z | pushed 2026-06-17T03:08:16Z
leandrofleury/plankton-defense
🛡️ 防范痞老板系统 | 🚷 守卫秘方的终极红队防线，大模型幻觉过滤、对抗攻击防御与不可逾越的合规死线。 | topics: estudo | stars: 0 | forks: 0 | updated 2026-06-18T12:45:07Z | pushed 2026-06-12T07:33:29Z
fengyun-shaohua/AI-Red-Teaming-Agent
AI Red Teaming Agent - 大模型红队安全评测工具 | language: Python | stars: 0 | forks: 1 | updated 2026-06-22T03:39:11Z | pushed 2026-06-22T03:39:07Z
shirley-dot/AI-test-llm-eval
Prompt 评测、用例生成、幻觉检测、红队检测，全部自动化 | language: Python | stars: 0 | forks: 0 | updated 2026-06-23T15:05:44Z | pushed 2026-06-23T15:05:34Z
AXFCVD/hongdui-gongfang-shizhan-jiaocai
红队攻防实战教材 | 18章系统化红队攻防知识体系：建→管→攻→防→应→测→拓 | stars: 6 | forks: 0 | updated 2026-07-21T07:47:33Z | pushed 2026-06-21T15:29:54Z
YuJi-89/fin_llm_implicit_memory_guard
本项目针对垂直金融大语言模型在微调过程中对量化策略、风控参数等隐性知识的记忆与泄露风险，设计并实现了一套检测与防护系统。系统主要包含两个核心模块：一是离线审计引擎，利用相对对数似然比检验（LiRT）与Min-K%方法量化评估模型的资产泄露风险，并执行自动化红队对抗压测；二是运行时通过轻量级NER、局部困惑度（PPL）以及采样语义熵，拦截硬编码记忆内容的输出。项目配套Streamlit前端界面，提供直观的数字化合规审计仪表盘。 | lan...