---
title: "agentic-engineering-principles"
description: "Principles for reliable agent workflows, tooling, verification, debugging, and guardrails"
canonical: https://agentpluginsdirectory.com/plugins/agentic-engineering-principles
last-updated: 2026-09-25
---

# agentic-engineering-principles
Principles for reliable agent workflows, tooling, verification, debugging, and guardrails
- Slug: agentic-engineering-principles
- Publisher: amhuppert
- Repository: https://github.com/amhuppert/my-ai-resources
- Manifest: plugins/agentic-engineering-principles/plugin.json
- Version: 1.3.0
- Category (editorial): other
- Skills: 13 (agent-feedback-tiers, agent-offloading, agent-retrospectives, agent-structured-output, ai-readable-tool-output, cli-tools-for-agents, designed-friction, earned-guidance-docs, live-system-verification, logging-for-agent-debugging, mechanical-guardrails, progressive-disclosure-tooling, query-output-disclosure)
- MCP servers: 0
- Stars: 2
- Repository created: 2025-07-06
- Repository last pushed: 2026-09-23
- Publisher type: User
- Listing: https://agentpluginsdirectory.com/plugins/agentic-engineering-principles
- Schema: https://agent-plugins.org/schemas/1.0.0/plugin.schema.json

## Skills

- agent-feedback-tiers: Define a three-tier feedback vocabulary: hint, reminders,
- agent-offloading: Covers the agent-offloading principle for AI-assisted development
- agent-retrospectives: Run structured retrospectives on any agent setup, instructions
- agent-structured-output: Covers how to get reliable machine-readable output from AI agents
- ai-readable-tool-output: "Covers configuring development tooling: linters, type checkers,
- cli-tools-for-agents: Design a purpose-built project CLI as the primary tool surface for
- designed-friction: Make a tool's deliberate constraints legible to agents, one-line
- earned-guidance-docs: "Applies an evidence-earned discipline to agent-facing
- live-system-verification: Verify features against the live running system, real user flows,
- logging-for-agent-debugging: Design structured logging that AI agents can actually debug from
- mechanical-guardrails: Converts recurring "please don't" guidance into structural
- progressive-disclosure-tooling: Structure a CLI or tool surface as a progressive-disclosure graph
- query-output-disclosure

Descriptions come from the frontmatter of each SKILL.md, punctuation lightly normalized.