roli-lpci/little-canary
Sacrificial LLM instances as behavioral probes for prompt injection detection
Combines structural pattern detection (regex + encoding-aware decode checks) with behavioral analysis via a small sacrificial LLM probe that runs at temperature=0 to detect compromise signals like persona adoption or instruction leakage. Supports both local Ollama models and OpenAI-compatible cloud APIs (MiniMax, Groq, Together, etc.), offering three detection modes (block, advisory, full) that integrate directly into existing LLM apps with ~250ms latency overhead and optional LLM-based verdict classification.
Available on PyPI.
Stars
7
Forks
—
Language
Python
License
Apache-2.0
Category
Last pushed
Mar 09, 2026
Monthly downloads
382
Commits (30d)
0
Dependencies
1
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/agents/roli-lpci/little-canary"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
Nebulock-Inc/agentic-threat-hunting-framework
ATHF is a framework for agentic threat hunting - building systems that can remember, learn, and...
AgentSeal/agentseal
Security toolkit for AI agents. Scan your machine for dangerous skills and MCP configs, monitor...
cosai-oasis/secure-ai-tooling
The CoSAI Risk Map is a framework for identifying, analyzing, and mitigating security risks in...
HeadyZhang/agent-audit
Static security scanner for LLM agents — prompt injection, MCP config auditing, taint analysis....
oasm-platform/open-asm
Open-source platform for cybersecurity Attack Surface Management (OASM).