- name
- strategic-compact
- description
- Context management intelligence for Claude Code sessions. ACTIVATE when context filling up, phase transitions detected, user mentions "running out of context" or "conversation too long". Suggests strategic compaction timing - manual compact at phase transitions beats auto-compact at arbitrary points.
# Strategic Compact
Intelligent context management that suggests compaction at optimal moments.
## Core Principle
**Manual compaction at strategic points > Auto-compaction at arbitrary points**
Auto-compact triggers at context limits, often mid-task. Strategic compaction preserves context through logical phases and compacts at natural transition points.
## When to Suggest Compaction
### Optimal Compaction Points
| Trigger | Reason |
|---------|--------|
| Exploration complete, implementation starting | Exploration context rarely needed for coding |
| Milestone completed | Clean slate for next phase |
| Plan finalized and documented | Plan captured, context can reset |
| Debug session resolved | Debug traces clutter future work |
| Switching to unrelated task | Previous context not relevant |
| 50+ tool calls in session | Accumulated context likely stale |
### Avoid Compaction During
| Scenario | Why |
|----------|-----|
| Mid-implementation | Loses code context and decisions |
| Active debugging | Loses diagnostic information |
| Incomplete task | May need earlier context |
| Code review in progress | Loses review thread |
## Compaction Options
### /compact (Quick)
- Built-in summarization
- Frees context immediately
- May lose some nuance
### /handoff (Recommended for complex work)
Creates HANDOFF.md with:
- Current goal and progress
- What worked / what didn't
- Next steps
- Key decisions made
Then start fresh with: `> HANDOFF.md`
### /clear (Fresh start)
- Complete reset
- Best when switching tasks entirely
## Phase Detection Patterns
Detect phase transitions by monitoring:
```
EXPLORATION → IMPLEMENTATION
- Many Read/Grep/Glob calls → Edit/Write calls starting
- Questions answered → Code being written
IMPLEMENTATION → TESTING
- Edit/Write heavy → Bash(test) calls
- Feature code done → Verification starting
DEBUGGING → RESOLUTION
- Error investigation → Fix confirmed working
- Multiple attempts → Success achieved
```
## Automatic Behavior
This skill works with hooks that:
1. **Track tool usage** - Counts tool calls per session
2. **Detect phase transitions** - Monitors tool patterns
3. **Inject suggestions** - Adds context-aware recommendations
## Usage Patterns
### After Exploration
```
[Hook detects: 30+ Read/Grep calls, first Edit coming]
→ Suggest: "Exploration complete. Consider /compact before implementation."
```
### After Milestone
```
[Hook detects: Tests passing after implementation]
→ Suggest: "Milestone reached. Good time for /compact or /handoff."
```
### Context Threshold
```
[Hook detects: 50 tool calls reached]
→ Suggest: "Session has 50+ tool calls. Consider /compact if context feels stale."
```
## Integration
Works automatically via plugin hooks. No manual configuration needed.
The hooks:
- Run on PreToolUse to track and suggest
- Run on PreCompact to add custom instructions
- Provide non-blocking suggestions via stderr
Auf GitHub ansehen