cy-workflow-memory

Read, update, promote, and compact workflow-scoped memory files for Compozy runs.

Updated May 15, 2026
One-click install
npx skills add https://github.com/JBonfim/skill-developer --skill cy-workflow-memory-jbonfim
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cy-workflow-memory
Source: https://github.com/JBonfim/skill-developer/tree/main/.agents/skills/cy-workflow-memory
Command: npx skills add https://github.com/JBonfim/skill-developer --skill cy-workflow-memory-jbonfim

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

It prevents workflow context from being lost between Compozy PRD task executions, so the agent can avoid repeated rediscovery and make consistent decisions across runs.

Core Features & Use Cases

  • Workflow-scoped memory maintenance: Load and update caller-provided workflow memory and current task memory files in the correct order.
  • Safe promotion of durable context: Promote only cross-task, durable facts into shared workflow memory using explicit criteria.
  • Noise control via compaction: Compact memory files when requested, guided by shared memory vs task memory roles to keep future runs clean.

Use case: During a multi-step delivery, a discovery about a cross-cutting constraint (e.g., rate limits or schema version dependency) is recorded once in shared workflow memory, while task-specific implementation details stay in the task-local file.

Quick Start

Ask the agent to run cy-workflow-memory with the provided workflow memory directory, shared workflow memory path, and current task memory path, and to update and compact the files as indicated by the caller signal.

Frequently Asked Questions about cy-workflow-memory

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

FAQPage Schema
How do I preserve workflow context across multi-run task executions?

To preserve workflow context across multi-run task executions, use a memory maintenance skill that reads, updates, and compacts shared memory files, ensuring durable cross-task facts carry forward while keeping task-local details isolated.

What is cross-task state management for multi-step agent workflows?

Cross-task state management for multi-step agent workflows is the process of promoting durable facts into shared memory files while isolating task-specific operational details, preventing noisy or incorrect history from contaminating future runs.

How do I compact memory files to control noise in workflow memory?

To compact memory files and control noise in workflow memory, execute a compaction process guided by the distinct roles of shared workflow memory versus task-local memory, keeping future runs clean and focused.

When should I promote task-local context into shared workflow memory?

You should promote task-local context into shared workflow memory when a discovery involves cross-cutting constraints, such as rate limits or schema version dependencies, using explicit criteria to ensure only durable facts are shared.

Do I need specific memory paths to run Compozy workflow memory maintenance?

Yes, you need specific memory paths to run Compozy workflow memory maintenance, as the process requires strict validation of the provided workflow memory directory, shared workflow memory path, and current task memory path.