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 h5i-dev · Codex Skill · ★ 54
Last updated: · Indexed by AgentSkillsHub · Auto-synced every 8h
🔒 Is awesome-ai-agent-incidents safe to install? View the security audit →
Awesome AI Agent Incidents A curated corpus of real-world security incidents, attack techniques, CVEs, frameworks, and defensive tools for autonomous AI agents. From zero-click Copilot exfiltration to AI-powered C2 channels — attacks on agentic AI have moved well beyond theory. Incidents like these raise a harder question for teams shipping AI-generated code: do you actually know what your agent did? We are building h5i to answer it — a Git sidecar that records every prompt, decision, and uncertainty signal alongside the diff, so you can audit what the AI touched, catch credential leaks before they land, and resume sessions without losing context. Table of Contents OWASP Agent Memory Guard – An official OWASP security framework for protecting AI agent memory from poisoning, injection, and exfiltration attacks. Provides detection middleware, sanitization hooks, and audit logging for LangChain, LlamaIndex, and custom agent pipelines. Awesome AI Agent Incidents Table of Contents Real-World Incidents
| Stars | 54 |
| Forks | 13 |
| Category | Codex Skill |
| License | MIT |
| Quality Score | 59.1308835516275/100 |
| Open Issues | 3 |
| Last Updated | 2026-09-30 |
| Created | 2026-03-30 |
| Platforms | claude-code, codex |
| Est. Tokens | ~28k |
Looking for a awesome-ai-agent-incidents alternative? If you're comparing awesome-ai-agent-incidents 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.
Open-source firewall for AI agents. Policy engine that audits and controls what OpenClaw, Claude Code, Cursor,
Jev demos, projects, SDKs and skills, with source links and a curated X gallery.
HeyClaude (formerly Claude Pro Directory) is a searchable collection of pre-built AI skills, agents, MCP serve
Dory treats hosted LLM interfaces as disposable tools by offloading state to your local environment and enforc
A collection of AI agent skills for Clawdbot, Claude Code, Codex
103 ready-to-use AI agent skills for Claude Code, OpenAI Codex, Gemini CLI, Cursor, GitHub Copilot, Windsurf,
Explore other popular codex skill tools:
awesome-ai-agent-incidents is A curated corpus of incidents, attack vectors, failure modes, and defensive tools for autonomous AI agents.. It is categorized as a Codex Skill with 54 GitHub stars.
You can find installation instructions and usage details in the awesome-ai-agent-incidents GitHub repository at github.com/h5i-dev/awesome-ai-agent-incidents. The project has 54 stars and 13 forks, indicating an active community.
awesome-ai-agent-incidents is released under the MIT license, making it free to use and modify according to the license terms.
The top alternatives to awesome-ai-agent-incidents on Agent Skills Hub include rampart, awesome-jev, claudepro-directory. 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: