skills.rest
josix
josixCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

agent-behavior-constraints

Define model routing, tool permissions, and behavioral guardrails for multi-agent orchestration.

7|1|Updated Feb 2, 2026
Search Tags:#agent orchestration#model routing#MCP tools#verification gates#behavioral guardrails#tool permissions#codex co-review
One-click install
npx skills add https://github.com/josix/agent-flow --skill agent-behavior-constraints
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-behavior-constraints
Source: https://github.com/josix/agent-flow/tree/main/skills/agent-behavior-constraints
Command: npx skills add https://github.com/josix/agent-flow --skill agent-behavior-constraints

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill prevents multi-agent orchestration failures by defining exactly how each persona must select models, use (or avoid) tools, and follow non-negotiable behavioral guardrails.

Core Features & Use Cases

  • Model routing: Assigns Opus vs Sonnet by agent role (planning/exploration vs execution/review/verification) to control speed/cost without sacrificing reasoning quality.
  • Tool access permissions: Enforces a per-agent tool matrix (read/grep/glob/write/edit/bash/web/task tools) including scoped exceptions for the authoring agent.
  • Behavioral guardrails: Standardizes safety rules like never speculating about unread code, avoiding suppressed type errors, preferring existing patterns, and requiring verification output completeness.
  • MCP and CLI preference handling: Requires domain operations to prefer MCP tools over shell/API fallbacks, and constrains any external CLI dispatch (Codex co-review) to a sanctioned orchestrator-only flow.
  • Violation protocol: Provides a deterministic response for forbidden actions: Stop, Document, Delegate, and Continue with permitted work.

Quick Start

Use the agent-behavior-constraints skill when orchestrating agents to ensure correct model selection, tool permissions, and guardrail enforcement for the current task.

Frequently Asked Questions about agent-behavior-constraints

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I enforce tool permissions and behavioral guardrails in a multi-agent orchestration workflow?▼

To enforce behavioral guardrails in multi-agent orchestration, define a per-agent tool access matrix that specifies read, write, and bash permissions for each persona, alongside universal non-negotiable rules to prevent forbidden operations.

What is the best way to route models like Opus and Sonnet across different agent roles?▼

The best way to route models for agent roles is assigning Opus to planning and exploration tasks for deeper reasoning, while dispatching Sonnet to execution, review, and verification tasks to optimize speed and cost.

Can I use MCP tools for domain operations while constraining external CLI dispatch like Codex co-review?▼

Yes, you can use MCP tools for domain operations by enforcing a preference over shell or API fallbacks, while constraining any external CLI dispatch such as Codex co-review to a sanctioned orchestrator-only flow.

What non-negotiable safety rules should I standardize for AI agent code generation and review?▼

Standardize AI agent safety rules by requiring agents to never speculate about unread code, avoid suppressing type errors, prefer existing code patterns, and ensure verification output completeness during review.

How do I handle a violation of behavioral guardrails during multi-agent orchestration?▼

When a multi-agent workflow violates behavioral guardrails, trigger a deterministic violation protocol that halts the forbidden operation, documents the issue, delegates the task to the correct agent, and continues with permitted work.

Related Skills

Explore More Software Engineering Skills→
JaviMontanoJaviMontano

metodologia-genai-architecture

Design GenAI architecture pipelines for RAG, model routing, and agent workflows.

Community
Advanced
bahayonghangbahayonghang
16

codex-bridge

Coordinate AI agent collaboration with Codex CLI for code implementation and review.

Community
Advanced
kevinnguyen271090kevinnguyen271090
15

mcp-management

Orchestrate MCP tool execution and discovery across multiple servers.

Community
Advanced
csuzngjhcsuzngjh
1

inject-rule

Appends user-provided rule arguments to the Ad-hoc Rules section of USER_CONTEXT.md.

Community
Basic