Pinned Loading

1. llm-jailbreak-techniques llm-jailbreak-techniques Public

   When an LLM refuses. 33 jailbreak techniques: obfuscation, prefill injection, roleplay framing.

   Python 2

2. ai-skill-security-scanner ai-skill-security-scanner Public

   Security audit for AI skills before installation. Catch credential leaks, suspicious patterns, and unsafe dependencies.

   1

3. vibe-safe vibe-safe Public

   Security pre-flight for AI vibe-coding agents. Audit dependencies before you code — catch CVEs when changing is still free.

   Shell

4. hermes-token-loop-postmortem hermes-token-loop-postmortem Public

   400M tokens burned by an AI agent loop. What happened, how to prevent it, and the cost comparison that will make you rethink your AI provider.

   1

5. agent-constitution-framework agent-constitution-framework Public

   Pre-action planning protocol for AI agents. PHASE 0-4 forces think-before-execute.