gaebalai

Aigis Kr — AI skill for Claude Code

AI community

Deterministic, zero-dependency Python firewall for AI agents — MCP rug-pull, memory poisoning, indirect injection, exfil channels.

How to install Aigis Kr

This entry records only its repository, not the path inside it, so there is no exact command to give. Open gaebalai/aigis-kr and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Aigis Kr does

Deterministic, zero-dependency Python firewall for AI agents — MCP rug-pull, memory poisoning, indirect injection, exfil channels. 44 compliance templates (US/CN/JP/EU).

Alternatives in AI

WindsurfAPI — Turn Windsurf / Devin Desktop's 100+ AI models (Claude, GPT, Gemini, DeepSeek, Kimi, GLM, SWE) into OpenAI-, A 3k ★
Image Use — Use your ChatGPT subscription to generate images from the command line — no OPENAI_API_KEY, no gateway, no dae 373 ★
ChatGPT Imagegen — Use your ChatGPT subscription to generate images from the command line — no OPENAI_API_KEY, no gateway, no dae 333 ★

README

Aigis

AI agents need agent-era security.
Chatbot-era guards filter text in and out. Aigis catches what comes next — MCP rug-pull at runtime, memory poisoning across sessions, capability escalation in tool calls, indirect injection in retrieved content.
Open source · zero dependencies · deterministic · no data leaves your environment.

Drop-in for Claude Code, Cursor, FastAPI, LangChain — 30 seconds via pip install pyaigis && aigis init.

100%
Detection on
paper-grounded
categories (76/76)

1,434
Tests Passing
on v1.1.0

44
Compliance Templates
(US/CN/JP/EU)

$0
Forever

Overall benchmark detection: 93.5% (144/154) with 0.0% false-positive rate (0/26 benign inputs). The 10 misses sit in alignment-frontier categories (sandbox escape, self-privilege escalation, audit tampering, evaluation gaming, CoT deception) — tracked on the L6/L7 verifier roadmap, not claimed as solved. See v1.1.0 release notes →