playbooks

home / skills / 89jobrien / steve / context-management

context-management skill

/steve/skills/context-management

This skill helps manage context across multi-agent workflows by capturing decisions, compressing data, and distributing state for coherent long-running

npx playbooks add skill 89jobrien/steve --skill context-management

Review the files below or copy the command above to add this skill to your agents.

Files (3)
SKILL.md
2.5 KB
---
name: context-management
description: Context management specialist for multi-agent workflows and long-running
  tasks. Use when coordinating multiple agents, preserving context across sessions,
  or managing complex project state. Focuses on context compression and distribution.
author: Joseph OBrien
status: unpublished
updated: '2025-12-23'
version: 1.0.1
tag: skill
type: skill
---

# Context Management

This skill manages context across multi-agent workflows and long-running tasks, ensuring coherent state preservation and efficient context distribution.

## When to Use This Skill

- When coordinating multiple agents in a workflow
- When preserving context across sessions
- When managing long-running projects
- When compressing context for efficiency
- When distributing context to specific agents
- When creating context checkpoints

## What This Skill Does

1. **Context Capture**: Extracts key decisions and rationale
2. **Context Distribution**: Prepares relevant context for each agent
3. **Memory Management**: Stores critical project decisions
4. **Context Compression**: Reduces context size while preserving key information
5. **Context Indexing**: Creates indexes for quick retrieval
6. **State Preservation**: Maintains coherent state across sessions

## How to Use

### Manage Context

```
Capture the key context from this conversation
```

```
Prepare context for the next agent in this workflow
```

### Specific Tasks

```
Create a context checkpoint for this project milestone
```

## Context Formats

### Quick Context (< 500 tokens)

- Current task and immediate goals
- Recent decisions affecting current work
- Active blockers or dependencies

### Full Context (< 2000 tokens)

- Project architecture overview
- Key design decisions
- Integration points and APIs
- Active work streams

### Archived Context

- Historical decisions with rationale
- Resolved issues and solutions
- Pattern library
- Performance benchmarks

## Best Practices

### Context Management

1. **Relevance Over Completeness**: Include only what's needed
2. **Key Decisions**: Capture rationale, not just outcomes
3. **Compression**: Reduce context while preserving meaning
4. **Indexing**: Create searchable context indexes
5. **Checkpoints**: Save context at major milestones

## Reference Files

- **`references/patterns.md`** - Context management patterns, handoff templates, and multi-agent coordination

## Related Use Cases

- Multi-agent coordination
- Long-running project management
- Context preservation across sessions
- Workflow state management
- Project knowledge management

Overview

This skill manages context across multi-agent workflows and long-running tasks to keep state coherent and actionable. It focuses on capturing decisions, compressing context, and distributing the right information to the right agents. Use it to preserve project memory, create checkpoints, and reduce context size without losing meaning.

How this skill works

The skill extracts key decisions, rationale, blockers, and metadata from conversations or project artifacts. It compresses and indexes that information, then prepares tailored context packages for individual agents or workflow stages. It also stores checkpoints and retrieves preserved state to maintain continuity across sessions.

When to use it

  • Coordinating multiple agents in a workflow
  • Preserving context across sessions or handoffs
  • Managing long-running projects and milestones
  • Compressing context for token- or bandwidth-limited systems
  • Creating and restoring context checkpoints

Best practices

  • Prioritize relevance: include only information needed for the next step
  • Capture rationale, not just outcomes, so future agents understand why
  • Compress context but preserve key decision traces and blockers
  • Index context for fast retrieval by agent, topic, or milestone
  • Save checkpoints at major milestones and before risky operations

Example use cases

  • Prepare a compact context bundle for the next agent in a handoff
  • Create a project checkpoint before a deployment or architecture change
  • Compress a long conversation into a <500-token quick context for fast decisions
  • Index and retrieve historical decisions when onboarding a new agent or team member
  • Distribute role-specific context subsets to agents in a multi-agent pipeline

FAQ

How do you decide what to compress versus what to keep intact?

Keep explicit decisions, blockers, and rationale intact; compress repetitive logs or low-value details. Use relevance scoring to prioritize items for full retention.

Can the skill restore context mid-project if sessions disconnect?

Yes. The skill stores checkpoints and indexed state so it can reconstruct the relevant context and resume workflows from the last saved point.