# Strategic Compact

> Guide for when to use /compact strategically. Use when context is getting large, when transitioning between task phases (research→planning→implementation), or when the user asks about compaction timing. Also use when tool call count is high or context feels stale.

- Skill: `kitfunso/strategic-compact` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add kitfunso/strategic-compact`
- Raw SKILL.md: https://api.skillmd.com/api/skills/kitfunso/strategic-compact/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: kitfunso (https://skillmd.com/u/kitfunso)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/kitfunso/strategic-compact

---


# Strategic Compact

Compact at **phase transitions**, not arbitrary token thresholds. The goal is to shed bulk while preserving what matters for the next phase.

## When to Compact

| Transition | Compact? | Why |
|---|---|---|
| Research → Planning | YES | Research is bulky; the plan is the distilled output |
| Planning → Implementation | YES | Plan is saved to file/TodoWrite; free context for code |
| After completing a milestone | YES | Fresh start for the new phase |
| Mid-implementation | NO | Losing variable names, paths, and reasoning is costly |
| Debugging → Next feature | YES | Clear dead-end traces and stack traces |
| Before context shift | YES | Clear exploration before switching to different task |

## What Survives Compaction
- CLAUDE.md instructions
- Memory files
- Git state awareness
- TodoWrite task lists

## What Gets Lost
- Intermediate reasoning and exploration
- File contents you read (must re-read)
- Conversation context and corrections
- Tool call history and error traces

## Best Practice
1. **Save your plan** to a file or TodoWrite BEFORE compacting
2. Compact with a descriptive message: `/compact Completed research phase. Plan saved to TODO. Starting implementation.`
3. After compacting, re-read any files you'll need for the next phase