Skip to main content

Du Dum: Batch Then Act Pattern

Skill Verified Active

Separate expensive observation from cheap decision-making in autonomous agent loops using a two-clock architecture. A fast clock accumulates data into a digest file; a slow clock reads the digest and acts only when something is pending. Idle cycles cost nothing because the action clock returns immediately after reading an empty digest. Use when building autonomous agents that must observe continuously but can only afford to act occasionally, when API or LLM costs dominate and most cycles have nothing to do, when designing cron-based agent architectures with observation and action phases, or when an existing heartbeat loop is too expensive because it calls the LLM on every tick.

Purpose

To provide a clear architectural pattern for building cost-efficient autonomous agents that need to observe continuously but act only occasionally, particularly when LLM or API costs are a significant concern.

Features

  • Two-clock architecture for observation and action separation
  • Digest file for low-bandwidth communication between clocks
  • Cost optimization for autonomous agent loops
  • Guidance on designing cron-based agent workflows
  • Handling of idle cycles with minimal cost

Use Cases

  • Building autonomous agents on a budget with continuous observation
  • Reducing costs in agent loops that frequently call LLMs or APIs unnecessarily
  • Designing decoupled agent systems with distinct analysis and action phases
  • Implementing efficient cron-based agent architectures

Non-Goals

  • Providing concrete executable code for the agent loop
  • Implementing specific LLM summarization or response generation logic
  • Handling real-time event processing that requires sub-second action

Workflow

  1. Identify the two clocks (observation and action)
  2. Design the digest format (compact, human-readable, machine-parseable)
  3. Implement the fast clock (analysis scripts)
  4. Implement the slow clock (action script)
  5. Configure idle detection for minimal overhead
  6. Validate the cost model for savings

Installation

/plugin install agent-almanac@pjt222-agent-almanac

Quality Score

Verified
96 /100
Analyzed about 21 hours ago

Trust Signals

Last commit1 day ago
Stars14
LicenseMIT
Status
View Source

Similar Extensions

Cloud Architect

100

Designs cloud architectures, creates migration plans, generates cost optimization recommendations, and produces disaster recovery strategies across AWS, Azure, and GCP. Use when designing cloud architectures, planning migrations, or optimizing multi-cloud deployments. Invoke for Well-Architected Framework, cost optimization, disaster recovery, landing zones, security architecture, serverless design.

Skill
jeffallan

Autonomous Agent Patterns

95

Design patterns for building autonomous coding agents. Covers tool integration, permission systems, browser automation, and human-in-the-loop workflows. Use when building AI agents, designing tool APIs, implementing permission systems, or creating autonomous coding assistants.

Skill
davila7

V3 Ddd Architecture

100

Domain-Driven Design architecture for claude-flow v3. Implements modular, bounded context architecture with clean separation of concerns and microkernel pattern.

Skill
ruvnet

Pathfinder

100

Map a codebase into feature-grouped flowcharts, identify duplicated concerns across features, and propose a unified architecture. Use when asked to "find the ideal path," unify duplicated systems, or audit architecture before a refactor. Emits a proposed unified flowchart plus per-system /make-plan prompts.

Skill
thedotmack

Understand

100

Analyze a codebase to produce an interactive knowledge graph for understanding architecture, components, and relationships

Skill
Lum1104

API Design Patterns

100

Comprehensive API design patterns covering REST, GraphQL, gRPC, versioning, authentication, and modern API best practices

Skill
bobmatnyc

© 2025 SkillRepo · Find the right skill, skip the noise.