grep - Oh My OpenCode

Overview

The grep tool provides fast content search using ripgrep with regex pattern matching. Use this when you need to find files containing specific patterns. Source: src/tools/grep/

Implementation

Uses ripgrep (rg) with:

60 second timeout
256KB output limit
Full regex syntax support
Multiple output modes
File pattern filtering

Parameters

pattern

string

required

The regex pattern to search for in file contentsExamples:

"function authenticate" - Literal string
"log.*Error" - Regex pattern
"function\s+\w+\(" - Function declarations
"TODO|FIXME" - Multiple patterns
"^import .* from" - Line-anchored pattern

Syntax: Full regex support (Rust regex)

. - Any character
* - Zero or more
+ - One or more
? - Zero or one
^ - Line start
$ - Line end
\s - Whitespace
\w - Word character
[abc] - Character class
(a|b) - Alternation

include

string

File pattern to include in the searchExamples:

"*.js" - JavaScript files only
"*.{ts,tsx}" - TypeScript files
"test/**/*.js" - Test files only

Uses glob pattern syntax.

path

string

The directory to search inDefaults to current working directory.Example: "src/components"

output_mode

enum

Output format

"files_with_matches" - Only file paths (default, fastest)
"content" - Matching lines with context
"count" - Match counts per file

Default: "files_with_matches"

head_limit

number

Limit output to first N entries0 or omitted means no limit.Example: head_limit: 10 returns first 10 matches.

Response

output

string

Search results in requested formatfiles_with_matches mode:

Found 5 files containing "authenticate":

src/auth/login.ts
src/auth/session.ts
src/middleware/auth.ts
tests/auth.test.ts
docs/authentication.md

content mode:

Found 8 matches in 3 files:

src/auth/login.ts
45: export function authenticate(username: string, password: string) {
67:   const token = await authenticateUser(credentials);

src/auth/session.ts
23: async function authenticateSession(token: string) {

count mode:

Match counts for "authenticate":

src/auth/login.ts: 12 matches
src/auth/session.ts: 5 matches
tests/auth.test.ts: 8 matches

Total: 25 matches in 3 files

No matches:

No matches found for "nonexistent"

Error:

Error: invalid regex pattern

Output modes

files_with_matches (default)

Fastest mode. Returns only file paths:

grep(pattern="TODO", output_mode="files_with_matches")

Use when you just need to know which files contain the pattern.

content

Returns matching lines with line numbers:

grep(pattern="function authenticate", output_mode="content")

Use when you need to see the actual matches.

count

Returns match counts per file:

grep(pattern="TODO", output_mode="count")

Use for statistics and metrics.

Usage examples

Find files containing pattern

# Default mode (fastest)
grep(pattern="authenticate")

# Explicit files_with_matches
grep(pattern="authenticate", output_mode="files_with_matches")

Search with file filter

# Only TypeScript files
grep(pattern="interface", include="*.ts")

# Only test files
grep(pattern="describe\(", include="*.test.{ts,js}")

View matching lines

# See actual matches
grep(pattern="function authenticate", output_mode="content")

# Limit to first 10 matches
grep(pattern="TODO", output_mode="content", head_limit=10)

Count matches

# Count TODOs
grep(pattern="TODO", output_mode="count")

# Count matches in specific files
grep(pattern="FIXME", include="*.ts", output_mode="count")

Search specific directory

grep(pattern="import", path="src/components")

Regex patterns

# Function declarations
grep(pattern="function\s+\w+\(")

# Import statements
grep(pattern="^import .* from")

# TODOs or FIXMEs
grep(pattern="TODO|FIXME")

# Error handling
grep(pattern="catch\s*\([^)]+\)\s*\{")

Common patterns

Find imports

# All imports
grep(pattern="^import")

# Specific package
grep(pattern="from ['\"]react['\"]")

# Default imports
grep(pattern="^import \w+ from")

Find function definitions

# Function keyword
grep(pattern="function\s+\w+")

# Arrow functions
grep(pattern="const\s+\w+\s*=\s*\([^)]*\)\s*=>")

# Async functions
grep(pattern="async\s+function")

Find TODO/FIXME comments

# All TODOs
grep(pattern="TODO", output_mode="count")

# TODOs with context
grep(pattern="TODO:", output_mode="content")

# FIXMEs in TypeScript
grep(pattern="FIXME", include="*.ts")

Find error handling

# Try-catch blocks
grep(pattern="try\s*\{")

# Error throwing
grep(pattern="throw new")

# Error logging
grep(pattern="console.error|logger.error")

Find API endpoints

# Express routes
grep(pattern="app\.(get|post|put|delete)")

# Fetch calls
grep(pattern="fetch\(")

# Axios calls
grep(pattern="axios\.(get|post)")

Safety limits

Timeout

Searches timeout after 60 seconds:

Error: Timeout after 60000ms

Refine your pattern or search a smaller directory.

Output limit

Output is limited to 256KB (about 3000 lines):

[Output truncated at 256KB. Refine your search for full results.]

Use head_limit or more specific patterns.

Head limit

Limit results to first N entries:

# First 10 files
grep(pattern="TODO", head_limit=10)

# First 5 matches with content
grep(pattern="function", output_mode="content", head_limit=5)

Useful for large result sets.

Ripgrep integration

The tool uses ripgrep with mode-specific flags: files_with_matches:

rg --files-with-matches pattern

content:

rg --line-number pattern

count:

rg --count pattern

Ripgrep is fast because it:

Uses memory maps for fast I/O
Respects .gitignore automatically
Skips binary files
Supports parallel search

Regex syntax

Ripgrep uses Rust regex syntax (similar to PCRE):

Character classes

\d - Digit
\w - Word character
\s - Whitespace
\D, \W, \S - Negations
[abc] - Character set
[^abc] - Negated set

Quantifiers

* - 0 or more
+ - 1 or more
? - 0 or 1
{n} - Exactly n
{n,} - n or more
{n,m} - Between n and m

Anchors

^ - Start of line
$ - End of line
\b - Word boundary

Groups

(pattern) - Capture group
(?:pattern) - Non-capture group
(a|b) - Alternation

Examples

# Email addresses
grep(pattern="\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Z|a-z]{2,}\b")

# URLs
grep(pattern="https?://[^\s]+")

# Hex colors
grep(pattern="#[0-9a-fA-F]{6}")

# Function calls
grep(pattern="\w+\([^)]*\)")

Case sensitivity

By default, searches are case-sensitive:

# Case-sensitive (matches "TODO" only)
grep(pattern="TODO")

# Case-insensitive (matches "todo", "TODO", "Todo")
grep(pattern="(?i)todo")

Use (?i) prefix for case-insensitive search.

Error handling

Invalid regex

Error: regex parse error:
    unclosed
    ^
unclosed group

Check your regex syntax:

Unmatched parentheses
Invalid escape sequences
Unclosed brackets

Ripgrep not found

Error: ripgrep not found. Install ripgrep or add it to PATH.

Install ripgrep (see glob tool).

Permission denied

Error: Permission denied

Check directory permissions or search a different path.

Performance tips

Use file filters

Instead of:

grep(pattern="import")

Use:

grep(pattern="import", include="*.ts")

Use files_with_matches for discovery

Instead of:

grep(pattern="authenticate", output_mode="content")

First:

grep(pattern="authenticate")  # Fast discovery

Then:

grep(pattern="authenticate", output_mode="content", path="src/auth")  # Focused search

Limit output

Instead of:

grep(pattern="function", output_mode="content")

Use:

grep(pattern="function", output_mode="content", head_limit=20)

Comparison with other tools

Feature	grep	grep (Unix)	ag	ast-grep
Speed	Very fast	Slow	Fast	Fast
Regex	Full	Basic/Extended	Full	Structural
Gitignore	Auto	Manual	Auto	Manual
Limits	Yes	No	No	No
Syntax	Regex	Regex	Regex	AST patterns

For code search, use grep for speed and safety. For structural search (syntax-aware), use ast_grep.

Implementation details

CLI execution

Executes ripgrep via runRg() or runRgCount():

const result = await runRg({
  pattern: args.pattern,
  paths: [searchPath],
  globs: args.include ? [args.include] : undefined,
  context: 0,
  outputMode,
  headLimit,
})

Result formatting

Results are formatted by formatGrepResult() or formatCountResult():

Adds context headers
Formats line numbers
Handles empty results
Adds truncation warnings

Thread limiting

Ripgrep thread count is limited to prevent resource exhaustion.

glob - File pattern matching
ast_grep - Structural code search
read - Read file contents
bash - Execute shell commands

Agents

Tools

Features

​Overview

​Implementation

​Parameters

​Response

​Output modes

​files_with_matches (default)

​content

​count

​Usage examples

​Find files containing pattern

​Search with file filter

​View matching lines

​Count matches

​Search specific directory

​Regex patterns

​Common patterns

​Find imports

​Find function definitions

​Find TODO/FIXME comments

​Find error handling

​Find API endpoints

​Safety limits

​Timeout

​Output limit

​Head limit

​Ripgrep integration

​Regex syntax

​Character classes

​Quantifiers

​Anchors

​Groups

​Examples

​Case sensitivity

​Error handling

​Invalid regex

​Ripgrep not found

​Permission denied

​Performance tips

​Use file filters

​Use files_with_matches for discovery

​Limit output

​Comparison with other tools

​Implementation details

​CLI execution

​Result formatting

​Thread limiting

​Related tools

Build docs developers (and LLMs) love

Overview

Implementation

Parameters

Response

Output modes

files_with_matches (default)

content

count

Usage examples

Find files containing pattern

Search with file filter

View matching lines

Count matches

Search specific directory

Regex patterns

Common patterns

Find imports

Find function definitions

Find TODO/FIXME comments

Find error handling

Find API endpoints

Safety limits

Timeout

Output limit

Head limit

Ripgrep integration

Regex syntax

Character classes

Quantifiers

Anchors

Groups

Examples

Case sensitivity

Error handling

Invalid regex

Ripgrep not found

Permission denied

Performance tips

Use file filters

Use files_with_matches for discovery

Limit output

Comparison with other tools

Implementation details

CLI execution

Result formatting

Thread limiting

Related tools