temper-agent

Mediate agent actions through Temper with Cedar authorization policies.

46|11|Updated Feb 9, 2026
One-click install
npx skills add https://github.com/nerdsane/temper --skill temper-agent
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: temper-agent
Source: https://github.com/nerdsane/temper/tree/main/skills/temper-agent
Command: npx skills add https://github.com/nerdsane/temper --skill temper-agent

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill provides the foundational operating layer for agents, ensuring all state-changing actions are securely mediated through the Temper framework, preventing direct API calls or sandbox escapes.

Core Features & Use Cases

  • Governed Execution: All agent actions flow through Temper, enforcing authorization policies via Cedar.
  • Spec-Driven Development: Agents model needs as state machines and data models, which Temper verifies and deploys.
  • Evolutionary Loop: Unmet intents are automatically recorded, enabling agents to propose spec changes for human approval and system improvement.
  • Use Case: An agent needs to manage a deployment pipeline. It models the deployment states (Pending, Building, Live, Rolledback) and submits this spec to Temper. All subsequent deployment actions are then governed by this verified spec and Cedar policies.

Quick Start

Use the temper-agent skill to discover all loaded specs for the default tenant.

Frequently Asked Questions about temper-agent

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

FAQPage Schema
How do I enforce authorization policies for agent state-changing actions?

You can enforce authorization policies for agent actions by mediating all operations through a governed sandbox environment that validates state machine models and applies Cedar policy rules before execution.

What is spec-driven development for secure agent operations?

Spec-driven development for secure agent operations is the process of modeling agent behaviors as state machines and data models, submitting them to a framework for formal verification, and governing all subsequent actions against the verified spec.

How do I prevent direct API calls or sandbox escapes from autonomous agents?

To prevent direct API calls or sandbox escapes, route all agent actions through an operating layer that mediates state changes, ensuring no execution occurs without formal verification and Cedar policy authorization.

Can I use Cedar policy to govern agent deployment pipelines?

Yes, you can use Cedar policy to govern agent deployment pipelines by modeling deployment states like Pending and Live as a state machine spec, then enforcing authorization policies on every state transition.

How do agents propose system improvements without unauthorized modifications?

Agents propose system improvements without unauthorized modifications by recording unmet intents in an evolutionary loop, automatically generating spec change proposals, and submitting them for human approval before any system modification occurs.

When should I not use a state machine approach for agent governance?

You should avoid using a state machine approach for agent governance when your operations require ad-hoc execution without formal verification, or when continuous human oversight for spec approval is not feasible.