grep-and-read

# Grep and Read Skill

**Purpose**: Search for pattern and read all matching files in a single coordinated operation, eliminating sequential round-trips.

**Performance**: 50-70% time savings, 4000-8000 token savings compared to sequential Grep -> Read -> Read -> Read pattern.

**When to Use**:
- Exploring codebase for specific functionality
- Finding all files containing a pattern
- Researching implementation approaches
- Investigating errors or bugs across multiple files

## Anti-Pattern This Skill Replaces

**INEFFICIENT PATTERN** (4 messages, 9-12 seconds):
```
Message 1: Grep "FormattingRule" -> returns 5 files
  [Wait for response ~2.5s]

Message 2: Read src/main/java/.../FormattingRule.java
  [Wait for response ~2.5s]

Message 3: Read src/test/java/.../FormattingRuleTest.java
  [Wait for response ~2.5s]

Message 4: Read docs/architecture.md
  [Wait for response ~2.5s]

-> Impact: 4 round-trips = ~10 seconds, ~12,000 tokens
-> Wasted: 3 avoidable round-trips
```

**EFFICIENT PATTERN** (1 message, 3-4 seconds):
```
Message 1: Skill grep-and-read pattern="FormattingRule" max_files=5
  - Grep finds matches
  - Read all matching files in parallel
  - Return consolidated output
  [Wait for response ~3.5s]

-> Impact: 1 round-trip = ~3.5 seconds, ~4,000 tokens
-> Saved: 3 round-trips = 6.5 seconds (65% faster), 8,000 tokens
```

## Skill Parameters

| Parameter | Required | Default | Description |
|-----------|----------|---------|-------------|
| `pattern` | Yes | - | Grep pattern to search for (regex supported) |
| `path` | No | `.` | Directory to search in |
| `glob` | No | - | File type filter (e.g., "*.java", "*.md") |
| `max_files` | No | 5 | Maximum number of files to read |
| `context_lines` | No | 100 | Lines to read per file (0 = all) |
| `case_sensitive` | No | true | Case-sensitive search |

## Skill Workflow

### Step 1: Search for Pattern

Use Grep tool to find all files containing the pattern:

```bash
# Search for pattern with optional filters
Grep: