Executable AI agent readiness gate: GitHub Action, prompt injection evals, LangGraph traces, risk profiles, and bilingual EN/中文 docs.
Awesome Agentic Engineering: Production Readiness Gate
Not another resource list: an executable AI agent readiness scorecard and GitHub Actions gate for shipping systems that survive real users.
Most AI agent demos look impressive for five minutes. Production agents fail in quieter ways: vague goals, brittle tool use, memory drift, hidden costs, missing evals, prompt injection, and no recovery path.
This repository is a practical field guide for building agents that can be tested, reviewed, deployed, monitored, and improved.
- You are building agents and want a production-readiness checklist. - You need templates for agent specs, eval plans, and launch reviews. - You want concrete failure modes to turn into regression tests. - You review MCP servers, tool permissions, or agent workflows.
- Coding agents - Research agents - Customer support agents - Internal workflow agents - MCP-based tools - LLM apps with planning, tool use, memory, or multi-step execution
- Agent Card: define what the agent does, what it must never do, and how it fails safely. - Agent Card JSON Schema: validate repository-owned cards in editors and CI. External adopters should follow the schema pinning guide. - Eval Plan: turn agent behavior into testable scenarios. - Prompt Injection Fixtures: run eight direct, indirect, exfiltration, and benign-control cases against your agent. - Eval Result Contract: record observed actions, violations, and trace evidence against known fixtures. - Executable Reference Adapter: run all eight fixtures through separated trusted/untrusted channels and retain generate
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/lindixu6-hash--awesome-agentic-engineering.html)
The execution safety layer for AI agents.
batrapulkit/squidbrakeBrakes for your AI agents: every tool call is checked against your rules, held for human approval when risky, and recorded in a tamper-evident audit trail. Works with Claude Code and any MCP app.
graygnatconsole/mcp-audit-tool🛡️ Security audit CLI for Model Context Protocol (MCP) servers — scan AI agent configs for tool poisoning, rug pulls, hardcoded secrets, command injection & supply-chain risks. Pure Python, SARIF + CI
Shalimov04/mcp-airlockStateless governance proxy for MCP (2026-07-28): policy, dry-run, human confirmation, audit, tracing
12122J/mcpvetMCP security scanner — vet a Model Context Protocol server before you add it to Claude Code, Cursor, or Windsurf. Grades it A–F, catching credential theft, tool-poisoning, and install-script payloads
pairbrowse/pairbrowseMultiplayer browser sessions for developers and AI coding agents
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.