> ## Documentation Index
> Fetch the complete documentation index at: https://condor.hummingbot.org/llms.txt
> Use this file to discover all available pages before exploring further.

# Managing Agent Sessions

> Monitor, inspect, and control running agents

Each time you start a Trading Agent, it creates a **session** that tracks all activity, decisions, and state changes. Here's how to manage them.

## View Active Sessions

```
/agent → Select agent → View Sessions

📂 SOL Scalper Sessions

Active:
• session_3 (running) - Started 2h ago, 120 ticks

Completed:
• session_2 - 6h, 360 ticks, +$45.20
• session_1 - 4h, 240 ticks, +$12.50
```

## Monitor a Running Session

```
/agent → SOL Scalper → View Status

🤖 SOL Scalper - Session 3
Status: Running
Tick: 121
Uptime: 2h 1m

Portfolio:
- Exposure: $75.00 / $100.00 limit
- Session P&L: +$23.40
- Active Executors: 3

Last Decision (Tick 121):
"Price consolidating at $156.20. EMAs still bullish.
 Holding current positions, waiting for breakout."
```

## View Session Journal

The journal contains the agent's complete decision history:

```
/agent → SOL Scalper → Journal

📓 Session 3 Journal

## Summary
- Started: 2024-03-27 14:00 UTC
- Current Tick: 121
- Session P&L: +$23.40

## Recent Decisions

### Tick 121 - 16:01 UTC
Decision: Hold
Reasoning: Price consolidating, waiting for breakout

### Tick 120 - 16:00 UTC
Decision: Take partial profit
Reasoning: Position #2 hit 1.5% gain, scaling out 50%
Action: Closed 5 SOL @ $157.80

### Tick 115 - 15:55 UTC
Decision: Open long
Reasoning: Price bounced off S1 ($154.50), EMAs aligned
Action: Created Position Executor - LONG 10 SOL @ $155.20
```

## View Snapshots

Snapshots capture the complete state of each tick for debugging:

```
/agent → SOL Scalper → Snapshots → Tick 115

📸 Snapshot - Tick 115

Prompt:
[Full system prompt sent to LLM]

Response:
"Analyzing SOL-USDT... Price at $155.20, bounced from
 support at $154.50. EMAs: 7 > 25 > 99 (bullish).
 Opening long position with 1% stop, 2% take profit."

Tool Calls:
1. run_routine("technical_analysis") → {trend: "bullish"...}
2. manage_executors("create", "position_executor") → {id: "exec_115"}

Risk State:
- Exposure: $50.00
- Daily P&L: +$18.20
- Is Blocked: false
```

## Inject Directives

Send real-time instructions to a running agent:

```
/agent → SOL Scalper → Inject Directive

Enter directive:
"Reduce exposure - high volatility expected in 10 minutes"

✓ Directive injected. Will appear in next tick.
```

The agent sees your directive and responds:

```
Tick 122 Decision:
"Received user directive about high volatility.
 Closing position #3 to reduce exposure from $75 to $50."
```

## Pause and Resume

**Pause** an agent without losing state:

```
/agent → SOL Scalper → Pause

⏸️ Agent paused. Executors remain active.
Resume anytime with /agent → Resume.
```

**Resume** continues from where it left off:

```
/agent → SOL Scalper → Resume

▶️ Agent resumed. Next tick in 45 seconds.
```

## View Learnings

Learnings persist across sessions:

```
/agent → SOL Scalper → Learnings

📚 SOL Scalper Learnings

## Active Insights

1. SOL tends to respect round numbers ($150, $155, $160)
2. Avoid entries during funding rate settlement (every 8h)
3. Wider stops (1.5%) work better in high volatility
4. EMAs lag during sharp moves - confirm with volume

## Session Notes

- Session 3: Consolidation day, fewer setups
- Session 2: Strong trend, 4/5 winners
```

## Stop and Archive

**Stop** ends the session:

```
/agent → SOL Scalper → Stop

🛑 Agent stopped.
Session 3 complete: 121 ticks, +$23.40

Open positions remain active. Manage manually or start new session.
```

Sessions are automatically archived and can be reviewed anytime.

## Next Step

<Card title="Routines" icon="arrow-right" href="/routines/overview">
  Extend agents with deterministic Python code
</Card>