pattern-detect

1# Pattern Detect
2 
3Identify recurring patterns using the Agent Monitor's workflow intelligence engine.
4 
5## Input
6 
7The user provides: **$ARGUMENTS**
8 
9Options: "all", "tools", "errors", "workflows", "last N sessions".
10 
11## Data Sources
12 
13| Endpoint | Returns |
14|----------|---------|
15| `GET /api/sessions?limit=200` | Session list with status, model, cwd, metadata |
16| `GET /api/analytics` | tool_usage top 20, event_types, agent_types |
17| `GET /api/workflows/{sessionId}` | 11 datasets per session (see below) |
18 
19### Workflow datasets used for pattern detection
20 
21| Dataset | Pattern insight |
22|---------|----------------|
23| `toolFlow` | **Tool transition matrix**: tool A → tool B with counts — reveals sequential habits |
24| `patterns` | **Detected workflow patterns**: recurring sequences with frequency scores |
25| `cooccurrence` | **Agent co-occurrence**: which agents frequently run together |
26| `modelDelegation` | **Model habits**: which models are chosen for which task types |
27| `errorPropagation` | **Error patterns**: where errors start and how they cascade by agent depth |
28| `effectiveness` | **Subagent patterns**: which types succeed most, avg duration per type |
29| `compaction` | **Compaction triggers**: what causes context overflow |
30| `complexity` | **Complexity patterns**: session complexity scores over time |
31 
32## Pattern Categories
33 
34### 1. Tool Chain Patterns (from `toolFlow`)
35- **Most common sequences**: Top 10 tool transitions (e.g., Read → Edit: 145 times)
36…