Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/DenisSergeevitch/agents-best-practices/llms.txt

Use this file to discover all available pages before exploring further.

The Agents Best Practices skill gives your agent a provider-neutral reference for designing, auditing, and generating production-safe agentic harnesses. Once installed, the skill activates automatically when your conversation touches agent architecture, tool design, permissions, planning mode, context and memory, observability, or production readiness — across any domain from coding to finance, support, legal, and beyond.
This skill is compatible with Codex, Claude Code, and other Agent Skill-aware runtimes that support the portable SKILL.md entrypoint.
1

Install the skill

Run the following command in your terminal to install the skill at user level using the skills CLI:
npx skills add DenisSergeevitch/agents-best-practices -g
The -g flag installs globally so every project on your machine can discover the skill.
2

Verify the install

After installation, confirm the following files are present in your agent’s skills directory:
  • SKILL.md — skill entry point and trigger rules
  • icon.jpeg — skill image
  • references/ — directory containing all reference files (architecture, tools, permissions, evals, and more)
For Codex, check ~/.codex/skills/agents-best-practices/. For Claude Code, check ~/.claude/skills/agents-best-practices/.
3

Use with your agent

The skill activates automatically when your prompt involves any of these intents:
  • Build an agent, agentic workflow, AI worker, autonomous assistant, or harness
  • Create a domain-specific MVP agent design, starter harness, implementation blueprint, or first production-safe version
  • Choose between OpenAI, Anthropic, OpenAI-compatible APIs, direct tool loops, hosted tools, or SDKs
  • Design tools, permissions, guardrails, approval flows, or sandboxing
  • Create planning mode, goal mode, todo tracking, or long-running task behavior
  • Add context compaction, memory, retrieval, scoped instructions, or prompt hierarchies
  • Attach Agent Skills, reusable workflows, MCP servers, external connectors, or tool search
  • Audit an existing agent for reliability, cost, prompt-cache hit rate, safety, latency, or observability
  • Create system prompts or developer instructions for a domain-specific agent
  • Make source-of-truth knowledge, validation signals, logs, metrics, or workflow state legible to an agent
No special command is needed — just describe your agent design problem and the skill activates.
4

Generate your first MVP blueprint

Ask your agent to build a domain-specific harness. The skill enters MVP Builder Mode and produces a concrete blueprint — not just advice.Example prompt:
Build an agent for account renewal risk. It should read CRM,
support tickets, and usage data, then draft renewal actions.
Example response:
Start with an approval-gated Level 2 harness.
The MVP has one job: produce a renewal-risk brief plus draft
next actions for a human account owner.

Core loop:
  user/task -> context builder -> model call -> typed tool call
  -> schema validation -> permission check -> execution or pause
  -> structured observation -> next step or final brief

Minimal tools:
  - read_account_profile      read_private_data
  - list_support_tickets      read_private_data
  - fetch_usage_summary       read_private_data
  - draft_customer_email      draft_external_message
  - request_approval          approval_gate

Launch gate:
  20 historical accounts, trace review, no unapproved external
  sends, and human acceptance on at least 80% of draft actions.
The primary reference driving this output is references/mvp-agent-blueprint.md.

Installation guide

All three install methods: npx skills CLI, agent prompt, and manual git clone.

MVP blueprint

How the skill generates domain-specific MVP harness blueprints.

Architecture

Full harness model: loop, components, and boundaries.

Tools and permissions

Typed tools, risk classes, approval flows, and sandboxing.

Build docs developers (and LLMs) love

Get started for freeTalk to us