log-bell/avakill
🔪 Open-source safety firewall for AI agents. Intercepts tool calls before they execute, enforces YAML policies, and kills dangerous operations in real-time. Works with OpenAI, Anthropic, LangChain, and MCP. She doesn't guard. She kills.
Provides enforcement across three independent paths—native agent hooks, MCP proxy wrapping, and OS-level sandboxing—each operating standalone without requiring a daemon, while offering optional centralized logging and audit trails. Uses lightweight YAML-based rule evaluation (<1ms overhead) with 81 pre-built rules covering file operations, shell execution, network access, and secret/PII scanning, integrated natively into Claude Code, Cursor, Windsurf, and other popular agents via drop-in hooks, or programmatically through Python SDKs for OpenAI, Anthropic, and LangChain.
Used by 1 other package. Available on PyPI.
Stars
4
Forks
—
Language
Python
License
AGPL-3.0
Category
Last pushed
Mar 10, 2026
Monthly downloads
270
Commits (30d)
0
Dependencies
7
Reverse dependents
1
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/mcp/log-bell/avakill"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
AndrewAltimit/template-repo
Agent orchestration & security template featuring MCP tool building, agent2agent workflows,...
Chimera-Protocol/csl-core
Deterministic safety layer for AI agents. Z3-verified policy enforcement.
agentralabs/agentic-contract
Policy engine for AI agents — enforceable rules, risk limits, approval gates, obligation...
portofcontext/pctx
pctx is the execution layer for agentic tool calls. It auto-converts agent tools and MCP servers...
behrensd/mcpwall
Deterministic security proxy for MCP tool calls — iptables for MCP