Implement comprehensive Ghidra Rust CLI with universal query system

This commit implements a complete Rust CLI tool for Ghidra reverse engineering, optimized for Claude Code and AI agents. Core Features: - Universal query command supporting all Ghidra data types (functions, strings, imports, exports, memory, etc.) - Advanced filter language with comparison, string, and logical operators - Multiple output formats (JSON, CSV, Table, minimal) optimized for LLM token efficiency - Field selection and pagination for precise data extraction - Windows-first design with cross-platform compatibility Architecture: - Filter parser using Pest grammar for robust expression parsing - Modular design with separate filter, format, query, and Ghidra integration layers - Headless Ghidra integration with built-in Python scripts for data extraction - Configuration system with environment variable and file support - Auto-detection of Ghidra installation on Windows LLM Optimizations: - Count-first workflow to check result sizes before fetching - Aggressive server-side filtering to reduce data transfer - Field selection to minimize token usage - Compact output formats (json-compact, minimal, ids) - Pagination support for large datasets Documentation: - Comprehensive README with examples and troubleshooting - Claude skill document (CLAUDE_SKILL.md) for agent integration - Subagent markdown (SUBAGENT.md) for Task tool integration - Inline code documentation and examples Commands Implemented: - ghidra query <data-type> - Universal query interface - ghidra import/analyze - Binary import and analysis - ghidra fn/strings/mem - Specialized command shortcuts - ghidra dump - Export data (imports, exports, functions, strings) - ghidra decompile - Function decompilation - ghidra project - Project management - ghidra config - Configuration management - ghidra init/doctor/version - Setup and diagnostics Built-in Ghidra Scripts: - Function listing with call graphs - Decompilation - String extraction - Import/Export tables - Memory map - Cross-references - Program information The CLI is designed to be succinct and efficient, with commands like: ghidra query functions --program=malware.exe --filter="size>1000 AND name~crypt" --format=json-compact Windows Support: - Auto-detection of Ghidra installation - Path handling for both Unix and Windows styles - Support for .exe, .dll, .sys formats This implementation provides a powerful, token-efficient interface for binary analysis that integrates seamlessly with Claude Code and other AI agents.
2026-07-10 03:18:56 -07:00 · 2026-01-12 21:11:19 +00:00
parent 790ad7153d
commit f8710ace0a
19 changed files with 6622 additions and 2 deletions
@@ -0,0 +1,463 @@
+# Ghidra CLI - Claude Code Skill
+
+This skill enables Claude Code and other AI agents to efficiently reverse engineer binaries using Ghidra through a powerful CLI interface.
+
+## Quick Reference
+
+### Most Common Commands
+
+```bash
+# Count-first workflow (ALWAYS use this pattern)
+ghidra query functions --program=<binary> --count
+ghidra query functions --program=<binary> --filter="<expr>" --count
+ghidra query functions --program=<binary> --filter="<expr>" --fields=name,address --format=json-compact
+
+# Query data types
+ghidra query functions|strings|imports|exports|memory --program=<binary> [options]
+
+# Decompile
+ghidra decompile <addr|name> --program=<binary>
+
+# Dump data
+ghidra dump imports|exports|functions|strings --program=<binary> [options]
+```
+
+## Setup & Initialization
+
+```bash
+# First time setup
+ghidra init
+
+# Check installation
+ghidra doctor
+
+# Import a binary
+ghidra import <binary-path> --project=<project-name>
+
+# Quick analysis (import + analyze + summary)
+ghidra quick <binary-path>
+```
+
+## Universal Query Command
+
+The `query` command is your primary tool. It supports filtering, field selection, and multiple output formats.
+
+### Syntax
+
+```bash
+ghidra query <data-type> --program=<binary> [options]
+```
+
+### Data Types
+
+- `functions` - All functions
+- `strings` - String literals
+- `imports` - Imported functions
+- `exports` - Exported functions
+- `memory` - Memory regions
+- `symbols` - Symbol table
+- `xrefs` - Cross-references
+
+### Options
+
+```bash
+--filter="<expression>"     # Filter results
+--fields=<list>             # Select specific fields (comma-separated)
+--format=<format>           # Output format (json, json-compact, table, minimal, count)
+--limit=<n>                 # Max results
+--offset=<n>                # Skip first n results
+--sort=<field>              # Sort by field (prefix with - for descending)
+--count                     # Just return count
+```
+
+## Filter Language
+
+### Comparison Operators
+
+```bash
+field=value          # Exact match
+field!=value         # Not equal
+field>value          # Greater than
+field>=value         # Greater or equal
+field<value          # Less than
+field<=value         # Less or equal
+```
+
+### String Operators
+
+```bash
+field~pattern        # Contains (case-insensitive)
+field^pattern        # Starts with
+field$pattern        # Ends with
+field=~regex         # Regex match
+```
+
+### Logical Operators
+
+```bash
+expr AND expr        # Both conditions
+expr OR expr         # Either condition
+NOT expr             # Negation
+(expr)               # Grouping
+```
+
+### Special Operators
+
+```bash
+field EXISTS         # Field is present
+field IN [val1,val2] # One of values
+```
+
+## Essential Workflows
+
+### 1. Initial Binary Analysis
+
+```bash
+# Get summary
+ghidra summary --program=<binary>
+
+# Count functions
+ghidra query functions --program=<binary> --count
+
+# Count named functions
+ghidra query functions --program=<binary> --filter="NOT name^FUN_" --count
+
+# List named functions (minimal output)
+ghidra query functions --program=<binary> \
+  --filter="NOT name^FUN_" \
+  --fields=name,address,size \
+  --format=json-compact \
+  --limit=50
+```
+
+### 2. Finding Interesting Functions
+
+```bash
+# Large functions
+ghidra query functions --program=<binary> \
+  --filter="size>1000" \
+  --fields=name,address,size \
+  --sort=-size \
+  --limit=20
+
+# Functions with specific keywords
+ghidra query functions --program=<binary> \
+  --filter="name~crypt OR name~encrypt OR name~password" \
+  --fields=name,address \
+  --format=json-compact
+
+# Functions that call specific APIs
+ghidra query functions --program=<binary> \
+  --filter="calls~WinExec OR calls~CreateProcess" \
+  --format=json-compact
+```
+
+### 3. String Analysis
+
+```bash
+# Count strings
+ghidra query strings --program=<binary> --count
+
+# Find URLs
+ghidra query strings --program=<binary> \
+  --filter="value~http" \
+  --fields=value,address \
+  --format=minimal
+
+# Find long strings (potential paths/URLs)
+ghidra query strings --program=<binary> \
+  --filter="length>50" \
+  --format=json-compact \
+  --limit=20
+
+# Find specific keywords
+ghidra query strings --program=<binary> \
+  --filter="value~password OR value~key OR value~token" \
+  --format=json-compact
+```
+
+### 4. Import Analysis
+
+```bash
+# List all imports
+ghidra dump imports --program=<binary> --format=json-compact
+
+# Find suspicious imports
+ghidra query imports --program=<binary> \
+  --filter="name IN [CreateProcess,WinExec,ShellExecute,WriteFile,CreateRemoteThread]" \
+  --format=json-compact
+
+# Find crypto imports
+ghidra query imports --program=<binary> \
+  --filter="name~Crypt" \
+  --format=json-compact
+```
+
+### 5. Decompilation
+
+```bash
+# Decompile by address
+ghidra decompile 0x401000 --program=<binary>
+
+# Decompile by name
+ghidra decompile main --program=<binary>
+
+# Decompile with minimal output
+ghidra fn decompile 0x401000 --program=<binary> --format=compact
+```
+
+### 6. Cross-Reference Analysis
+
+```bash
+# Find what calls a function
+ghidra query xrefs --program=<binary> \
+  --filter="to~WinExec" \
+  --fields=from,from_function \
+  --format=json-compact
+
+# Find all callers to an address
+ghidra xref to 0x401000 --program=<binary> --format=json-compact
+```
+
+## Output Formats
+
+Choose the right format for your use case:
+
+- `count` - Just the number (best for checking result size)
+- `json-compact` - Minimal JSON (best for LLMs)
+- `minimal` - Just addresses/names (good for piping)
+- `ids` - Just IDs (good for further queries)
+- `table` - Human-readable table (good for display)
+- `json` - Full JSON (when you need all data)
+
+## Best Practices for LLMs
+
+### 1. Always Count First
+
+```bash
+# BAD: Fetching all data without knowing size
+ghidra query functions --program=<binary>
+
+# GOOD: Count first, then filter
+ghidra query functions --program=<binary> --count
+ghidra query functions --program=<binary> --filter="size>1000" --count
+ghidra query functions --program=<binary> --filter="size>1000" --format=json-compact
+```
+
+### 2. Use Aggressive Filtering
+
+```bash
+# BAD: Fetching then filtering in code
+ghidra query functions --program=<binary> --format=json
+
+# GOOD: Filter on Ghidra side
+ghidra query functions --program=<binary> \
+  --filter="size>1000 AND name~crypt" \
+  --format=json-compact
+```
+
+### 3. Select Only Needed Fields
+
+```bash
+# BAD: Getting all fields
+ghidra query functions --program=<binary>
+
+# GOOD: Select only what you need
+ghidra query functions --program=<binary> \
+  --fields=name,address,size \
+  --format=json-compact
+```
+
+### 4. Paginate Large Results
+
+```bash
+# Get first page
+ghidra query functions --program=<binary> --limit=50
+
+# Get next page
+ghidra query functions --program=<binary> --limit=50 --offset=50
+```
+
+### 5. Use Appropriate Output Format
+
+```bash
+# For analysis: json-compact
+ghidra query functions --program=<binary> --format=json-compact
+
+# For display: table
+ghidra query functions --program=<binary> --format=table
+
+# For piping: minimal or ids
+ghidra query functions --program=<binary> --format=ids
+```
+
+## Common Analysis Patterns
+
+### Pattern 1: Find Entry Points
+
+```bash
+# Find main or WinMain
+ghidra query functions --program=<binary> \
+  --filter="name~main OR name~WinMain OR name~DllMain" \
+  --format=json-compact
+```
+
+### Pattern 2: Find Crypto Functions
+
+```bash
+# By name
+ghidra query functions --program=<binary> \
+  --filter="name~crypt OR name~cipher OR name~hash OR name~aes OR name~rsa" \
+  --format=json-compact
+
+# By imports
+ghidra query imports --program=<binary> \
+  --filter="name~Crypt" \
+  --format=json-compact
+```
+
+### Pattern 3: Find Network Functions
+
+```bash
+# By imports
+ghidra query imports --program=<binary> \
+  --filter="name~socket OR name~connect OR name~send OR name~recv OR name~http" \
+  --format=json-compact
+```
+
+### Pattern 4: Find File Operations
+
+```bash
+ghidra query imports --program=<binary> \
+  --filter="name~File OR name~Read OR name~Write OR name~Create" \
+  --format=json-compact
+```
+
+### Pattern 5: Find Suspicious Strings
+
+```bash
+# Registry keys
+ghidra query strings --program=<binary> \
+  --filter="value~HKEY OR value~Software" \
+  --format=json-compact
+
+# URLs
+ghidra query strings --program=<binary> \
+  --filter="value~http" \
+  --format=json-compact
+
+# Credentials
+ghidra query strings --program=<binary> \
+  --filter="value~password OR value~username OR value~token" \
+  --format=json-compact
+```
+
+## Error Handling
+
+### Common Errors
+
+1. **Program not specified**: Use `--program=<binary>` or set default with `ghidra set-default program <binary>`
+2. **Ghidra not found**: Run `ghidra init` or set `GHIDRA_INSTALL_DIR`
+3. **Analysis timeout**: Increase with `set GHIDRA_TIMEOUT=600`
+4. **Project not found**: Create with `ghidra project create <name>`
+
+### Troubleshooting
+
+```bash
+# Check installation
+ghidra doctor
+
+# List available projects
+ghidra project list
+
+# Show current configuration
+ghidra config list
+```
+
+## Configuration
+
+### Set Defaults
+
+```bash
+# Set default program (so you don't have to pass --program each time)
+ghidra set-default program <binary>
+
+# Set default project
+ghidra set-default project <project-name>
+```
+
+### Environment Variables
+
+```bash
+# Windows
+set GHIDRA_INSTALL_DIR=C:\ghidra\ghidra_11.0
+set GHIDRA_DEFAULT_PROGRAM=malware.exe
+
+# Unix
+export GHIDRA_INSTALL_DIR=/opt/ghidra
+export GHIDRA_DEFAULT_PROGRAM=malware.elf
+```
+
+## Example Workflow
+
+Here's a complete analysis workflow:
+
+```bash
+# 1. Import and analyze
+ghidra import suspicious.exe --project=analysis
+
+# 2. Get overview
+ghidra summary --program=suspicious.exe
+
+# 3. Count functions
+ghidra query functions --program=suspicious.exe --count
+# Output: 1247
+
+# 4. Count named functions
+ghidra query functions --program=suspicious.exe --filter="NOT name^FUN_" --count
+# Output: 89
+
+# 5. Get named functions
+ghidra query functions --program=suspicious.exe \
+  --filter="NOT name^FUN_" \
+  --fields=name,address,size \
+  --format=json-compact
+
+# 6. Find suspicious imports
+ghidra dump imports --program=suspicious.exe \
+  --filter="name~Exec OR name~Create OR name~Write" \
+  --format=json-compact
+
+# 7. Find interesting strings
+ghidra query strings --program=suspicious.exe \
+  --filter="value~http OR value~password" \
+  --format=json-compact
+
+# 8. Decompile interesting functions
+ghidra decompile 0x401000 --program=suspicious.exe
+```
+
+## Tips
+
+1. **Always count before fetching** - Prevents overwhelming your context
+2. **Use filters aggressively** - Pre-filter on Ghidra side
+3. **Select minimal fields** - Reduces token usage
+4. **Use json-compact format** - Most efficient for LLMs
+5. **Set defaults** - Avoids repeating `--program` and `--project`
+6. **Paginate large results** - Use `--limit` and `--offset`
+7. **Cache results** - Store commonly-used queries in variables
+
+## Advanced: Chaining Commands
+
+```bash
+# Get list of function addresses, then decompile each
+FUNCS=$(ghidra query functions --program=<binary> \
+  --filter="name~suspicious" \
+  --format=ids)
+
+for addr in $FUNCS; do
+  ghidra decompile $addr --program=<binary> --format=compact
+done
+```
+
+This skill gives you powerful, token-efficient access to Ghidra for binary analysis!
@@ -0,0 +1,64 @@
+[package]
+name = "ghidra-cli"
+version = "0.1.0"
+edition = "2021"
+authors = ["Alexander Kiselev"]
+description = "Rust CLI to run Ghidra headless for reverse engineering with Claude Code and other agents"
+license = "GPL-3.0"
+repository = "http://127.0.0.1:62915/git/akiselev/ghidra-cli"
+
+[dependencies]
+# CLI framework
+clap = { version = "4.5", features = ["derive", "env", "cargo"] }
+
+# Serialization
+serde = { version = "1.0", features = ["derive"] }
+serde_json = "1.0"
+serde_yaml = "0.9"
+
+# Parsing
+pest = "2.7"
+pest_derive = "2.7"
+
+# Output formatting
+tabled = "0.15"
+comfy-table = "7.1"
+
+# Error handling
+anyhow = "1.0"
+thiserror = "1.0"
+
+# Logging
+env_logger = "0.11"
+log = "0.4"
+
+# File system & paths
+dirs = "5.0"
+tempfile = "3.8"
+walkdir = "2.4"
+
+# Process management
+which = "6.0"
+
+# Regex
+regex = "1.10"
+lazy_static = "1.4"
+
+# CSV/TSV
+csv = "1.3"
+
+# String similarity (for fuzzy matching)
+strsim = "0.11"
+
+# Cross-platform support
+dunce = "1.0"  # Windows path handling
+atty = "0.2"   # TTY detection
+
+[dev-dependencies]
+assert_cmd = "2.0"
+predicates = "3.0"
+tempfile = "3.8"
+
+[[bin]]
+name = "ghidra"
+path = "src/main.rs"
@@ -1,2 +1,425 @@
-# ghidra-cli
-Rust cli to run ghidra headless on files so that Claude Code and other agents can reverse engineer stuff
+# Ghidra CLI
+
+A powerful Rust CLI tool for Ghidra reverse engineering, designed for Claude Code and other AI agents to efficiently analyze binaries.
+
+## Features
+
+- 🚀 **Universal Query System** - Query any Ghidra data type with a single command
+- 🔍 **Advanced Filtering** - Powerful filter language for precise data extraction
+- 📊 **Multiple Output Formats** - JSON, CSV, TSV, Table, and more
+- 🤖 **LLM-Optimized** - Designed for minimal token usage and maximum efficiency
+- 🪟 **Windows-First** - Native Windows support with cross-platform compatibility
+- 📦 **Zero Configuration** - Auto-detection of Ghidra installation
+- ⚡ **Fast** - Direct headless Ghidra integration
+
+## Installation
+
+### Prerequisites
+
+- [Ghidra](https://ghidra-sre.org/) 10.0 or later
+- Rust 1.70+ (for building from source)
+
+### From Source
+
+```bash
+git clone https://github.com/yourusername/ghidra-cli
+cd ghidra-cli
+cargo build --release
+```
+
+The binary will be at `target/release/ghidra.exe` (Windows) or `target/release/ghidra` (Unix).
+
+### Setup
+
+1. Run the initialization wizard:
+```bash
+ghidra init
+```
+
+2. Set your Ghidra installation (if not auto-detected):
+```bash
+set GHIDRA_INSTALL_DIR=C:\ghidra\ghidra_11.0
+```
+
+3. Verify installation:
+```bash
+ghidra doctor
+```
+
+## Quick Start
+
+```bash
+# Quick analysis of a binary
+ghidra quick malware.exe
+
+# Import a binary
+ghidra import suspicious.exe --project=analysis
+
+# Query functions
+ghidra query functions --program=suspicious.exe --filter="size>1000"
+
+# Decompile a function
+ghidra decompile 0x401000 --program=suspicious.exe
+
+# List suspicious imports
+ghidra dump imports --program=suspicious.exe --filter="name~Crypt OR name~Process"
+```
+
+## Universal Query Command
+
+The `query` command is the primary interface for data extraction:
+
+```bash
+ghidra query <data-type> [options]
+```
+
+### Supported Data Types
+
+- `functions` - All functions in the program
+- `strings` - String data
+- `imports` - Import table
+- `exports` - Export table
+- `memory` - Memory regions
+- `symbols` - Symbol table
+- `xrefs` - Cross-references
+- `comments` - All comments
+- `types` - Data types
+
+### Query Options
+
+```bash
+--filter="<expression>"   # Filter results
+--fields=<list>           # Select specific fields
+--format=<format>         # Output format
+--limit=<n>               # Max results
+--offset=<n>              # Skip first n results
+--sort=<field>            # Sort order
+--count                   # Just return count
+```
+
+### Filter Language
+
+```bash
+# Comparison operators
+name=malloc               # Exact match
+size>1000                 # Greater than
+address>=0x401000         # Greater or equal
+
+# String operators
+name~crypt                # Contains (case-insensitive)
+name^sub_                 # Starts with
+name$_exit                # Ends with
+name=~"regex"             # Regex match
+
+# Logical operators
+name~crypt AND size>500   # AND
+name~main OR name~start   # OR
+NOT name^FUN_             # NOT
+
+# Special operators
+calls EXISTS              # Field exists
+name IN [malloc,free]     # One of values
+size>=100 AND size<=1000  # Range
+```
+
+## Examples
+
+### Function Analysis
+
+```bash
+# List all functions
+ghidra query functions --program=malware.exe
+
+# Find large functions with crypto in the name
+ghidra query functions --program=malware.exe \
+  --filter="size>1000 AND name~crypt" \
+  --format=json-compact
+
+# Count unnamed functions
+ghidra query functions --program=malware.exe \
+  --filter="name^FUN_" \
+  --count
+
+# Get specific fields only
+ghidra query functions --program=malware.exe \
+  --fields=name,address,size \
+  --limit=10
+```
+
+### String Analysis
+
+```bash
+# Find HTTP/HTTPS URLs
+ghidra query strings --program=malware.exe \
+  --filter="value~http"
+
+# Find long strings (potential paths, URLs)
+ghidra query strings --program=malware.exe \
+  --filter="length>50" \
+  --format=minimal
+```
+
+### Import Analysis
+
+```bash
+# Find suspicious imports
+ghidra query imports --program=malware.exe \
+  --filter="name IN [CreateProcess,WinExec,ShellExecute]"
+
+# Find crypto imports
+ghidra query imports --program=malware.exe \
+  --filter="name~Crypt" \
+  --format=table
+```
+
+### Memory Analysis
+
+```bash
+# List executable memory regions
+ghidra query memory --program=malware.exe \
+  --filter="permissions~x" \
+  --format=table
+```
+
+### Decompilation
+
+```bash
+# Decompile a specific function
+ghidra decompile 0x401000 --program=malware.exe
+
+# Decompile by name
+ghidra decompile main --program=malware.exe
+
+# Get compact output
+ghidra fn decompile suspicious_func --program=malware.exe \
+  --format=compact
+```
+
+## Specialized Commands
+
+### Function Commands
+
+```bash
+ghidra fn list [options]                    # List functions
+ghidra fn get <addr|name> [options]         # Get function details
+ghidra fn decompile <addr|name> [options]   # Decompile
+ghidra fn calls <addr|name> [options]       # What it calls
+ghidra fn xrefs <addr|name> [options]       # What calls it
+```
+
+### String Commands
+
+```bash
+ghidra strings [options]                    # List all strings
+ghidra strings refs <string> [options]      # Get references
+```
+
+### Memory Commands
+
+```bash
+ghidra mem map [options]                    # Memory map
+ghidra mem read <addr> <size> [options]     # Read memory
+ghidra mem search <pattern> [options]       # Search for pattern
+```
+
+### Dump Commands
+
+```bash
+ghidra dump imports [options]               # All imports
+ghidra dump exports [options]               # All exports
+ghidra dump functions [options]             # All functions
+ghidra dump strings [options]               # All strings
+```
+
+## Output Formats
+
+```
+full          - Full human-readable (default for TTY)
+compact       - One-line summaries
+minimal       - Just addresses/names
+json          - Full JSON
+json-compact  - Minimal JSON
+json-stream   - NDJSON (one per line)
+csv           - CSV format
+tsv           - TSV format
+table         - Pretty table
+ids           - Just addresses/IDs
+count         - Just count
+```
+
+## Configuration
+
+### Environment Variables
+
+```bash
+GHIDRA_INSTALL_DIR       # Path to Ghidra installation
+GHIDRA_PROJECT_DIR       # Project directory
+GHIDRA_DEFAULT_PROGRAM   # Default program to analyze
+GHIDRA_DEFAULT_PROJECT   # Default project name
+GHIDRA_TIMEOUT           # Command timeout (seconds)
+```
+
+### Configuration File
+
+Located at:
+- Windows: `%APPDATA%\ghidra-cli\config.yaml`
+- Linux/Mac: `~/.config/ghidra-cli/config.yaml`
+
+```yaml
+ghidra_install_dir: C:\ghidra\ghidra_11.0
+ghidra_project_dir: C:\Users\username\.ghidra-projects
+default_program: malware.exe
+default_project: analysis
+default_output_format: json-compact
+default_limit: 1000
+timeout: 300
+```
+
+### Set Defaults
+
+```bash
+# Set default program
+ghidra set-default program malware.exe
+
+# Set default project
+ghidra set-default project analysis
+```
+
+## LLM-Optimized Workflow
+
+For Claude Code and other agents:
+
+```bash
+# 1. Count first (check result size)
+ghidra query functions --program=malware.exe --count
+# → 1,247 functions
+
+# 2. Refine filter and count
+ghidra query functions --program=malware.exe \
+  --filter="NOT name^FUN_" \
+  --count
+# → 89 named functions
+
+# 3. Get minimal data
+ghidra query functions --program=malware.exe \
+  --filter="NOT name^FUN_" \
+  --fields=name,address \
+  --format=json-compact
+
+# 4. Deep dive on specific items
+ghidra fn decompile <address> --program=malware.exe \
+  --format=compact
+```
+
+## Project Management
+
+```bash
+# Create project
+ghidra project create myproject
+
+# List projects
+ghidra project list
+
+# Delete project
+ghidra project delete myproject
+```
+
+## Scripting
+
+### Run Custom Scripts
+
+```bash
+# Run a Python script
+ghidra script run my_analysis.py --program=malware.exe -- arg1 arg2
+
+# Execute inline Python
+ghidra script python "print(currentProgram.getName())" --program=malware.exe
+
+# Execute inline Java
+ghidra script java "println(currentProgram.getName());" --program=malware.exe
+```
+
+### Built-in Scripts
+
+The CLI includes built-in scripts for:
+- Function listing
+- Decompilation
+- String extraction
+- Import/Export tables
+- Memory map
+- Cross-references
+- Program information
+
+## Windows-Specific Notes
+
+### Path Handling
+
+The CLI handles both Unix-style (`/`) and Windows-style (`\`) paths automatically.
+
+### Ghidra Installation Detection
+
+Auto-detection checks these locations:
+- `C:\Program Files\Ghidra`
+- `C:\Program Files (x86)\Ghidra`
+- `C:\ghidra`
+- Registry entries (if available)
+
+### Executable Detection
+
+Supports common Windows formats:
+- `.exe` - Executables
+- `.dll` - Dynamic libraries
+- `.sys` - System drivers
+
+## Performance Tips
+
+1. **Use `--count` first** - Check result size before fetching data
+2. **Filter aggressively** - Pre-filter on Ghidra side, not in your code
+3. **Select minimal fields** - Use `--fields` to reduce data transfer
+4. **Use compact formats** - `json-compact` or `minimal` for LLMs
+5. **Paginate large results** - Use `--limit` and `--offset`
+
+## Troubleshooting
+
+### Ghidra Not Found
+
+```bash
+# Check doctor
+ghidra doctor
+
+# Set manually
+set GHIDRA_INSTALL_DIR=C:\path\to\ghidra
+
+# Or in config
+ghidra config set ghidra_install_dir C:\path\to\ghidra
+```
+
+### Analysis Timeout
+
+```bash
+# Increase timeout
+set GHIDRA_TIMEOUT=600
+
+# Or in config
+ghidra config set timeout 600
+```
+
+### Project Issues
+
+```bash
+# List projects
+ghidra project list
+
+# Delete and recreate
+ghidra project delete myproject
+ghidra project create myproject
+```
+
+## License
+
+GNU General Public License v3.0 - see [LICENSE](LICENSE) for details.
+
+## Credits
+
+- [Ghidra](https://ghidra-sre.org/) - NSA's reverse engineering framework
+- Built with ❤️ for Claude Code and the AI agent community
@@ -0,0 +1,483 @@
+# Ghidra Binary Analysis Subagent
+
+## Overview
+
+This subagent specializes in reverse engineering binaries using Ghidra CLI. It provides efficient, token-optimized access to binary analysis capabilities for Claude Code and other AI agents.
+
+## When to Use This Subagent
+
+Use this subagent when you need to:
+- Analyze binary executables (PE, ELF, Mach-O)
+- Reverse engineer malware or suspicious binaries
+- Extract functions, strings, imports/exports from binaries
+- Decompile functions to understand behavior
+- Find specific patterns in binary code
+- Analyze memory layout and structure
+- Identify crypto functions, network operations, or file I/O
+- Generate reports on binary capabilities
+
+## Capabilities
+
+### Data Extraction
+- **Functions**: List, filter, and decompile functions
+- **Strings**: Extract and search string literals
+- **Imports/Exports**: Analyze external dependencies
+- **Memory Layout**: View memory regions and permissions
+- **Symbols**: Access symbol table
+- **Cross-References**: Find call relationships
+
+### Analysis Features
+- **Universal Query System**: Query any data type with powerful filters
+- **Advanced Filtering**: Complex boolean expressions with field-level filtering
+- **Multiple Output Formats**: JSON, CSV, Table, minimal (token-efficient)
+- **Decompilation**: Convert assembly to C-like pseudocode
+- **Pattern Matching**: Find specific code patterns and strings
+
+### LLM Optimizations
+- **Count-First Workflow**: Check result sizes before fetching data
+- **Field Selection**: Request only needed fields
+- **Aggressive Filtering**: Pre-filter on Ghidra side
+- **Compact Formats**: Minimal token usage with `json-compact`
+- **Pagination**: Handle large datasets efficiently
+
+## Command Reference
+
+### Quick Start
+
+```bash
+# Import and analyze a binary
+ghidra import <binary-path> --project=<project>
+
+# Quick analysis (all-in-one)
+ghidra quick <binary-path>
+
+# Get program summary
+ghidra summary --program=<binary>
+```
+
+### Universal Query
+
+```bash
+# Query any data type
+ghidra query <data-type> --program=<binary> [options]
+
+# Data types: functions, strings, imports, exports, memory, symbols, xrefs
+
+# Essential options:
+  --filter="<expression>"    # Filter results
+  --fields=<list>            # Select specific fields
+  --format=<format>          # Output format (json, json-compact, table, count)
+  --limit=<n>                # Max results
+  --count                    # Just return count
+```
+
+### Common Queries
+
+```bash
+# List functions with filtering
+ghidra query functions --program=<binary> \
+  --filter="size>1000 AND name~crypt" \
+  --fields=name,address,size \
+  --format=json-compact
+
+# Find strings
+ghidra query strings --program=<binary> \
+  --filter="value~http" \
+  --format=minimal
+
+# List imports
+ghidra dump imports --program=<binary> \
+  --filter="name~Crypt" \
+  --format=json-compact
+
+# Get memory map
+ghidra query memory --program=<binary> --format=table
+```
+
+### Decompilation
+
+```bash
+# Decompile function by address
+ghidra decompile 0x401000 --program=<binary>
+
+# Decompile by name
+ghidra decompile main --program=<binary>
+
+# Compact output
+ghidra fn decompile <addr> --program=<binary> --format=compact
+```
+
+## Filter Language
+
+### Operators
+
+```
+Comparison:     =, !=, >, >=, <, <=
+String:         ~ (contains), ^ (starts), $ (ends), =~ (regex)
+Logical:        AND, OR, NOT, ()
+Special:        EXISTS, IN [val1,val2]
+```
+
+### Examples
+
+```bash
+# Exact match
+name=malloc
+
+# Numeric comparison
+size>1000
+
+# String matching (case-insensitive)
+name~crypt
+
+# Boolean logic
+name~crypt AND size>500
+(name~main OR name~start) AND NOT name^FUN_
+
+# IN operator
+name IN [malloc,free,realloc]
+
+# Field existence
+calls EXISTS
+
+# Complex expression
+size>=100 AND size<=1000 AND (name~crypt OR calls~Crypt)
+```
+
+## Output Formats
+
+- `count` - Just the number (check result size)
+- `json-compact` - Minimal JSON (best for LLMs)
+- `minimal` - Addresses/names only (piping)
+- `ids` - Just IDs (for further queries)
+- `table` - Human-readable (display)
+- `json` - Full JSON (complete data)
+
+## Best Practices
+
+### 1. Count-First Pattern
+
+Always check the result size before fetching data:
+
+```bash
+# Step 1: Count
+ghidra query functions --program=<binary> --count
+
+# Step 2: Refine filter if needed
+ghidra query functions --program=<binary> \
+  --filter="NOT name^FUN_" \
+  --count
+
+# Step 3: Fetch minimal data
+ghidra query functions --program=<binary> \
+  --filter="NOT name^FUN_" \
+  --fields=name,address \
+  --format=json-compact \
+  --limit=50
+```
+
+### 2. Aggressive Filtering
+
+Filter on Ghidra side, not in your code:
+
+```bash
+# GOOD: Pre-filter
+ghidra query functions --program=<binary> \
+  --filter="size>1000 AND name~crypt"
+
+# BAD: Fetch all, then filter
+ghidra query functions --program=<binary>  # Then filter in code
+```
+
+### 3. Field Selection
+
+Request only what you need:
+
+```bash
+# Only name and address
+ghidra query functions --program=<binary> \
+  --fields=name,address \
+  --format=json-compact
+```
+
+### 4. Use Compact Formats
+
+Minimize token usage:
+
+```bash
+# For analysis: json-compact
+--format=json-compact
+
+# For display: table
+--format=table
+
+# For piping: minimal or ids
+--format=ids
+```
+
+## Analysis Workflows
+
+### Initial Reconnaissance
+
+```bash
+# 1. Get summary
+ghidra summary --program=<binary>
+
+# 2. Count functions
+ghidra query functions --program=<binary> --count
+
+# 3. Count named functions
+ghidra query functions --program=<binary> \
+  --filter="NOT name^FUN_" --count
+
+# 4. List key functions
+ghidra query functions --program=<binary> \
+  --filter="NOT name^FUN_" \
+  --fields=name,address,size \
+  --format=json-compact \
+  --limit=20
+```
+
+### Finding Suspicious Behavior
+
+```bash
+# Network operations
+ghidra query imports --program=<binary> \
+  --filter="name~socket OR name~http OR name~inet" \
+  --format=json-compact
+
+# File operations
+ghidra query imports --program=<binary> \
+  --filter="name~File OR name~Read OR name~Write" \
+  --format=json-compact
+
+# Process operations
+ghidra query imports --program=<binary> \
+  --filter="name~Process OR name~Thread OR name~Exec" \
+  --format=json-compact
+
+# Crypto operations
+ghidra query imports --program=<binary> \
+  --filter="name~Crypt" \
+  --format=json-compact
+```
+
+### String Analysis
+
+```bash
+# URLs
+ghidra query strings --program=<binary> \
+  --filter="value~http" \
+  --format=json-compact
+
+# Registry keys
+ghidra query strings --program=<binary> \
+  --filter="value~HKEY OR value~Software" \
+  --format=json-compact
+
+# Credentials
+ghidra query strings --program=<binary> \
+  --filter="value~password OR value~username OR value~key" \
+  --format=json-compact
+
+# Long strings (paths, URLs)
+ghidra query strings --program=<binary> \
+  --filter="length>50" \
+  --format=json-compact
+```
+
+### Deep Dive Analysis
+
+```bash
+# 1. Find interesting functions
+ghidra query functions --program=<binary> \
+  --filter="name~crypt OR calls~Crypt" \
+  --fields=name,address \
+  --format=json-compact
+
+# 2. Decompile each
+ghidra decompile <address> --program=<binary>
+
+# 3. Find cross-references
+ghidra xref to <address> --program=<binary> \
+  --format=json-compact
+
+# 4. Analyze callers
+ghidra fn calls <address> --program=<binary> \
+  --format=json-compact
+```
+
+## Common Patterns
+
+### Pattern: Find Entry Point
+
+```bash
+ghidra query functions --program=<binary> \
+  --filter="name~main OR name~WinMain OR name~DllMain" \
+  --format=json-compact
+```
+
+### Pattern: Find Crypto Functions
+
+```bash
+# By name
+ghidra query functions --program=<binary> \
+  --filter="name~crypt OR name~cipher OR name~aes OR name~rsa" \
+  --format=json-compact
+
+# By imports
+ghidra query imports --program=<binary> \
+  --filter="name~Crypt" \
+  --format=json-compact
+```
+
+### Pattern: Find Large/Complex Functions
+
+```bash
+ghidra query functions --program=<binary> \
+  --filter="size>2000" \
+  --fields=name,address,size \
+  --sort=-size \
+  --limit=10 \
+  --format=json-compact
+```
+
+### Pattern: Trace Function Calls
+
+```bash
+# What does function call?
+ghidra fn calls <function> --program=<binary> \
+  --format=json-compact
+
+# What calls this function?
+ghidra fn xrefs <function> --program=<binary> \
+  --format=json-compact
+```
+
+## Configuration
+
+### Set Defaults (avoid repeating --program)
+
+```bash
+# Set default program
+ghidra set-default program <binary>
+
+# Now you can omit --program
+ghidra query functions --count
+```
+
+### Environment Variables
+
+```bash
+# Windows
+set GHIDRA_INSTALL_DIR=C:\ghidra\ghidra_11.0
+set GHIDRA_DEFAULT_PROGRAM=malware.exe
+
+# Unix
+export GHIDRA_INSTALL_DIR=/opt/ghidra
+export GHIDRA_DEFAULT_PROGRAM=malware.elf
+```
+
+## Error Handling
+
+### Common Issues
+
+1. **Ghidra not found**: Run `ghidra init` or set `GHIDRA_INSTALL_DIR`
+2. **Program not specified**: Use `--program=<binary>` or set default
+3. **Analysis timeout**: Increase with `set GHIDRA_TIMEOUT=600`
+4. **Large result set**: Use `--count` first, then filter more aggressively
+
+### Troubleshooting
+
+```bash
+# Check installation
+ghidra doctor
+
+# Show configuration
+ghidra config list
+
+# List projects
+ghidra project list
+```
+
+## Performance Considerations
+
+1. **Always count first** - Prevents context overflow
+2. **Filter aggressively** - Reduce data before transfer
+3. **Select minimal fields** - Less data = fewer tokens
+4. **Use compact formats** - `json-compact` is most efficient
+5. **Paginate results** - Use `--limit` for large datasets
+6. **Cache results** - Store commonly-used data in variables
+
+## Example: Complete Analysis
+
+```bash
+# Import binary
+ghidra import suspicious.exe --project=analysis
+
+# Set as default
+ghidra set-default program suspicious.exe
+
+# Overview
+ghidra summary
+
+# Count functions
+ghidra query functions --count
+# → 1247
+
+# Named functions only
+ghidra query functions --filter="NOT name^FUN_" --count
+# → 89
+
+# Get named functions
+ghidra query functions \
+  --filter="NOT name^FUN_" \
+  --fields=name,address,size \
+  --format=json-compact
+
+# Find suspicious imports
+ghidra dump imports \
+  --filter="name~Exec OR name~Process OR name~Write" \
+  --format=json-compact
+
+# Find URLs/IPs
+ghidra query strings \
+  --filter="value~http OR value=~\"[0-9]{1,3}\\.[0-9]{1,3}\"" \
+  --format=json-compact
+
+# Decompile interesting functions
+ghidra decompile 0x401000
+
+# Find what calls it
+ghidra xref to 0x401000 --format=json-compact
+```
+
+## Integration Tips
+
+This subagent works best when:
+- You have a binary file that needs analysis
+- You need to understand malware behavior
+- You're investigating suspicious executables
+- You need to extract specific information (strings, functions, imports)
+- You want to generate a report on binary capabilities
+
+The subagent is optimized for:
+- Token efficiency (minimal output)
+- Fast queries (count-first pattern)
+- Precise filtering (server-side pre-filtering)
+- Flexible output (multiple formats)
+- Automation-friendly (scriptable)
+
+## Limitations
+
+- Requires Ghidra to be installed
+- Windows path handling is primary (but cross-platform)
+- Initial analysis can be slow for large binaries
+- Decompilation quality depends on Ghidra's capabilities
+- Complex analysis may require multiple queries
+
+## Support
+
+- Run `ghidra doctor` to check installation
+- See `README.md` for full documentation
+- Check `CLAUDE_SKILL.md` for detailed examples
@@ -0,0 +1,171 @@
+use serde::{Deserialize, Serialize};
+use std::path::PathBuf;
+use std::fs;
+use crate::error::{GhidraError, Result};
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct Config {
+    pub ghidra_install_dir: Option<PathBuf>,
+    pub ghidra_project_dir: Option<PathBuf>,
+    pub default_program: Option<String>,
+    pub default_project: Option<String>,
+    pub default_output_format: Option<String>,
+    pub default_limit: Option<usize>,
+    pub timeout: Option<u64>,
+    pub aliases: std::collections::HashMap<String, String>,
+}
+
+impl Default for Config {
+    fn default() -> Self {
+        Self {
+            ghidra_install_dir: None,
+            ghidra_project_dir: None,
+            default_program: None,
+            default_project: None,
+            default_output_format: Some("auto".to_string()),
+            default_limit: Some(1000),
+            timeout: Some(300),
+            aliases: std::collections::HashMap::new(),
+        }
+    }
+}
+
+impl Config {
+    pub fn load() -> Result<Self> {
+        let config_path = Self::config_path()?;
+
+        if !config_path.exists() {
+            return Ok(Self::default());
+        }
+
+        let content = fs::read_to_string(&config_path)?;
+        let config: Config = serde_yaml::from_str(&content)?;
+
+        Ok(config)
+    }
+
+    pub fn save(&self) -> Result<()> {
+        let config_path = Self::config_path()?;
+
+        if let Some(parent) = config_path.parent() {
+            fs::create_dir_all(parent)?;
+        }
+
+        let content = serde_yaml::to_string(self)?;
+        fs::write(config_path, content)?;
+
+        Ok(())
+    }
+
+    pub fn config_path() -> Result<PathBuf> {
+        let config_dir = dirs::config_dir()
+            .ok_or_else(|| GhidraError::ConfigError("Could not determine config directory".to_string()))?;
+
+        Ok(config_dir.join("ghidra-cli").join("config.yaml"))
+    }
+
+    pub fn get_ghidra_install_dir(&self) -> Result<PathBuf> {
+        // Check environment variable first
+        if let Ok(dir) = std::env::var("GHIDRA_INSTALL_DIR") {
+            return Ok(PathBuf::from(dir));
+        }
+
+        // Check config
+        if let Some(dir) = &self.ghidra_install_dir {
+            return Ok(dir.clone());
+        }
+
+        // Try to auto-detect on Windows
+        #[cfg(target_os = "windows")]
+        {
+            if let Some(dir) = Self::detect_ghidra_windows() {
+                return Ok(dir);
+            }
+        }
+
+        Err(GhidraError::GhidraNotFound)
+    }
+
+    pub fn get_project_dir(&self) -> Result<PathBuf> {
+        // Check environment variable first
+        if let Ok(dir) = std::env::var("GHIDRA_PROJECT_DIR") {
+            return Ok(PathBuf::from(dir));
+        }
+
+        // Check config
+        if let Some(dir) = &self.ghidra_project_dir {
+            return Ok(dir.clone());
+        }
+
+        // Default to ~/.ghidra-projects
+        let home = dirs::home_dir()
+            .ok_or_else(|| GhidraError::ConfigError("Could not determine home directory".to_string()))?;
+
+        Ok(home.join(".ghidra-projects"))
+    }
+
+    #[cfg(target_os = "windows")]
+    fn detect_ghidra_windows() -> Option<PathBuf> {
+        // Check common installation paths
+        let common_paths = vec![
+            PathBuf::from("C:\\Program Files\\Ghidra"),
+            PathBuf::from("C:\\Program Files (x86)\\Ghidra"),
+            PathBuf::from("C:\\ghidra"),
+        ];
+
+        for path in common_paths {
+            if path.exists() {
+                // Look for ghidra_* directories
+                if let Ok(entries) = fs::read_dir(&path) {
+                    for entry in entries.flatten() {
+                        let entry_path = entry.path();
+                        if entry_path.is_dir() {
+                            let name = entry_path.file_name()?.to_str()?;
+                            if name.starts_with("ghidra_") {
+                                // Check if analyzeHeadless.bat exists
+                                let headless = entry_path.join("support").join("analyzeHeadless.bat");
+                                if headless.exists() {
+                                    return Some(entry_path);
+                                }
+                            }
+                        }
+                    }
+                }
+            }
+        }
+
+        None
+    }
+
+    pub fn get_timeout(&self) -> u64 {
+        std::env::var("GHIDRA_TIMEOUT")
+            .ok()
+            .and_then(|s| s.parse().ok())
+            .or(self.timeout)
+            .unwrap_or(300)
+    }
+
+    pub fn get_default_program(&self) -> Option<String> {
+        std::env::var("GHIDRA_DEFAULT_PROGRAM")
+            .ok()
+            .or_else(|| self.default_program.clone())
+    }
+
+    pub fn get_default_project(&self) -> Option<String> {
+        std::env::var("GHIDRA_DEFAULT_PROJECT")
+            .ok()
+            .or_else(|| self.default_project.clone())
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_default_config() {
+        let config = Config::default();
+        assert_eq!(config.timeout, Some(300));
+        assert_eq!(config.default_limit, Some(1000));
+    }
+}
@@ -0,0 +1,57 @@
+use thiserror::Error;
+
+#[derive(Error, Debug)]
+pub enum GhidraError {
+    #[error("Ghidra installation not found. Set GHIDRA_INSTALL_DIR or run 'ghidra init'")]
+    GhidraNotFound,
+
+    #[error("Ghidra project not found: {0}")]
+    ProjectNotFound(String),
+
+    #[error("Program not found: {0}")]
+    ProgramNotFound(String),
+
+    #[error("Failed to execute Ghidra: {0}")]
+    ExecutionFailed(String),
+
+    #[error("Failed to parse filter: {0}")]
+    FilterParseError(String),
+
+    #[error("Invalid filter expression: {0}")]
+    InvalidFilter(String),
+
+    #[error("Field not found: {0}")]
+    FieldNotFound(String),
+
+    #[error("Invalid format: {0}")]
+    InvalidFormat(String),
+
+    #[error("Invalid data type: {0}")]
+    InvalidDataType(String),
+
+    #[error("Configuration error: {0}")]
+    ConfigError(String),
+
+    #[error("IO error: {0}")]
+    IoError(#[from] std::io::Error),
+
+    #[error("JSON error: {0}")]
+    JsonError(#[from] serde_json::Error),
+
+    #[error("YAML error: {0}")]
+    YamlError(#[from] serde_yaml::Error),
+
+    #[error("Command failed: {0}")]
+    CommandFailed(String),
+
+    #[error("Invalid address: {0}")]
+    InvalidAddress(String),
+
+    #[error("Analysis timeout after {0} seconds")]
+    Timeout(u64),
+
+    #[error("{0}")]
+    Other(String),
+}
+
+pub type Result<T> = std::result::Result<T, GhidraError>;
@@ -0,0 +1,36 @@
+WHITESPACE = _{ " " | "\t" | "\n" | "\r" }
+
+// Main expression
+expr = { logical_expr }
+
+// Logical expressions
+logical_expr = { logical_term ~ (logical_op ~ logical_term)* }
+logical_term = { logical_not | "(" ~ logical_expr ~ ")" | comparison }
+logical_not = { ("NOT" | "!") ~ logical_term }
+logical_op = { "AND" | "OR" | "&&" | "||" }
+
+// Comparison expressions
+comparison = { field ~ compare_op ~ value | field ~ string_op ~ string_value | existence_check | in_check }
+
+// Operators
+compare_op = { ">=" | "<=" | "!=" | "=" | ">" | "<" }
+string_op = { "=~" | "~" | "^" | "$" }
+
+// Existence checks
+existence_check = { field ~ ("EXISTS" | "EMPTY" | "NULL") }
+
+// IN operator
+in_check = { field ~ "IN" ~ "[" ~ value_list ~ "]" }
+value_list = { value ~ ("," ~ value)* }
+
+// Field access (including nested fields and arrays)
+field = @{ identifier ~ ("." ~ identifier | "[" ~ number ~ "]")* }
+identifier = @{ (ASCII_ALPHA | "_") ~ (ASCII_ALPHANUMERIC | "_")* }
+
+// Values
+value = { number | hex_number | boolean | quoted_string | identifier }
+string_value = { quoted_string | identifier }
+number = @{ "-"? ~ ASCII_DIGIT+ ~ ("." ~ ASCII_DIGIT+)? }
+hex_number = @{ "0x" ~ ASCII_HEX_DIGIT+ }
+boolean = { "true" | "false" | "TRUE" | "FALSE" }
+quoted_string = @{ "\"" ~ (!"\"" ~ ANY)* ~ "\"" | "'" ~ (!"'" ~ ANY)* ~ "'" }
@@ -0,0 +1,256 @@
+use regex::Regex;
+use serde_json::Value as JsonValue;
+use crate::error::{GhidraError, Result};
+use super::{FilterExpr, CompareOp, StringOp, LogicalOp, ExistenceCheck, Value};
+
+pub fn evaluate(expr: &FilterExpr, data: &JsonValue) -> Result<bool> {
+    match expr {
+        FilterExpr::Compare { field, op, value } => {
+            evaluate_compare(field, *op, value, data)
+        }
+        FilterExpr::StringOp { field, op, value } => {
+            evaluate_string_op(field, *op, value, data)
+        }
+        FilterExpr::Logical { op, exprs } => {
+            evaluate_logical(*op, exprs, data)
+        }
+        FilterExpr::Not(inner) => {
+            Ok(!evaluate(inner, data)?)
+        }
+        FilterExpr::Exists { field, check } => {
+            evaluate_exists(field, *check, data)
+        }
+        FilterExpr::In { field, values } => {
+            evaluate_in(field, values, data)
+        }
+    }
+}
+
+fn get_field_value<'a>(field: &str, data: &'a JsonValue) -> Option<&'a JsonValue> {
+    let parts: Vec<&str> = field.split('.').collect();
+    let mut current = data;
+
+    for part in parts {
+        // Check for array index like "field[0]"
+        if let Some(bracket_pos) = part.find('[') {
+            let field_name = &part[..bracket_pos];
+            let index_str = &part[bracket_pos + 1..part.len() - 1];
+
+            current = current.get(field_name)?;
+
+            if let Ok(index) = index_str.parse::<usize>() {
+                current = current.get(index)?;
+            } else {
+                return None;
+            }
+        } else {
+            current = current.get(part)?;
+        }
+    }
+
+    Some(current)
+}
+
+fn evaluate_compare(field: &str, op: CompareOp, value: &Value, data: &JsonValue) -> Result<bool> {
+    let field_value = get_field_value(field, data);
+
+    if field_value.is_none() {
+        return Ok(false);
+    }
+
+    let field_value = field_value.unwrap();
+
+    match (field_value, value) {
+        (JsonValue::Number(n), val) => {
+            let field_num = n.as_f64().unwrap();
+            let compare_num = val.as_f64()
+                .ok_or_else(|| GhidraError::InvalidFilter(format!("Cannot compare number with {:?}", val)))?;
+
+            Ok(match op {
+                CompareOp::Equal => (field_num - compare_num).abs() < f64::EPSILON,
+                CompareOp::NotEqual => (field_num - compare_num).abs() >= f64::EPSILON,
+                CompareOp::Greater => field_num > compare_num,
+                CompareOp::GreaterEqual => field_num >= compare_num,
+                CompareOp::Less => field_num < compare_num,
+                CompareOp::LessEqual => field_num <= compare_num,
+            })
+        }
+        (JsonValue::String(s), Value::String(val)) => {
+            Ok(match op {
+                CompareOp::Equal => s == val,
+                CompareOp::NotEqual => s != val,
+                _ => return Err(GhidraError::InvalidFilter("Cannot use numeric comparison on strings".to_string())),
+            })
+        }
+        (JsonValue::Bool(b), Value::Boolean(val)) => {
+            Ok(match op {
+                CompareOp::Equal => *b == *val,
+                CompareOp::NotEqual => *b != *val,
+                _ => return Err(GhidraError::InvalidFilter("Cannot use numeric comparison on booleans".to_string())),
+            })
+        }
+        _ => Ok(false),
+    }
+}
+
+fn evaluate_string_op(field: &str, op: StringOp, value: &str, data: &JsonValue) -> Result<bool> {
+    let field_value = get_field_value(field, data);
+
+    if field_value.is_none() {
+        return Ok(false);
+    }
+
+    let field_str = match field_value.unwrap() {
+        JsonValue::String(s) => s.to_lowercase(),
+        JsonValue::Number(n) => n.to_string(),
+        JsonValue::Bool(b) => b.to_string(),
+        _ => return Ok(false),
+    };
+
+    let value_lower = value.to_lowercase();
+
+    Ok(match op {
+        StringOp::Contains => field_str.contains(&value_lower),
+        StringOp::StartsWith => field_str.starts_with(&value_lower),
+        StringOp::EndsWith => field_str.ends_with(&value_lower),
+        StringOp::Regex => {
+            let re = Regex::new(value)
+                .map_err(|e| GhidraError::InvalidFilter(format!("Invalid regex: {}", e)))?;
+            re.is_match(&field_str)
+        }
+    })
+}
+
+fn evaluate_logical(op: LogicalOp, exprs: &[FilterExpr], data: &JsonValue) -> Result<bool> {
+    match op {
+        LogicalOp::And => {
+            for expr in exprs {
+                if !evaluate(expr, data)? {
+                    return Ok(false);
+                }
+            }
+            Ok(true)
+        }
+        LogicalOp::Or => {
+            for expr in exprs {
+                if evaluate(expr, data)? {
+                    return Ok(true);
+                }
+            }
+            Ok(false)
+        }
+    }
+}
+
+fn evaluate_exists(field: &str, check: ExistenceCheck, data: &JsonValue) -> Result<bool> {
+    let field_value = get_field_value(field, data);
+
+    Ok(match check {
+        ExistenceCheck::Exists => field_value.is_some(),
+        ExistenceCheck::Empty => {
+            match field_value {
+                None => true,
+                Some(JsonValue::Null) => true,
+                Some(JsonValue::String(s)) => s.is_empty(),
+                Some(JsonValue::Array(a)) => a.is_empty(),
+                Some(JsonValue::Object(o)) => o.is_empty(),
+                _ => false,
+            }
+        }
+        ExistenceCheck::Null => {
+            matches!(field_value, None | Some(JsonValue::Null))
+        }
+    })
+}
+
+fn evaluate_in(field: &str, values: &[Value], data: &JsonValue) -> Result<bool> {
+    let field_value = get_field_value(field, data);
+
+    if field_value.is_none() {
+        return Ok(false);
+    }
+
+    let field_value = field_value.unwrap();
+
+    for val in values {
+        match (field_value, val) {
+            (JsonValue::String(s), Value::String(v)) => {
+                if s.eq_ignore_ascii_case(v) {
+                    return Ok(true);
+                }
+            }
+            (JsonValue::Number(n), v) => {
+                if let Some(compare_num) = v.as_f64() {
+                    if (n.as_f64().unwrap() - compare_num).abs() < f64::EPSILON {
+                        return Ok(true);
+                    }
+                }
+            }
+            (JsonValue::Bool(b), Value::Boolean(v)) => {
+                if *b == *v {
+                    return Ok(true);
+                }
+            }
+            _ => {}
+        }
+    }
+
+    Ok(false)
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use serde_json::json;
+
+    #[test]
+    fn test_evaluate_compare() {
+        let data = json!({
+            "name": "test",
+            "size": 100
+        });
+
+        let expr = FilterExpr::Compare {
+            field: "size".to_string(),
+            op: CompareOp::Greater,
+            value: Value::Integer(50),
+        };
+
+        assert!(evaluate(&expr, &data).unwrap());
+    }
+
+    #[test]
+    fn test_evaluate_string_op() {
+        let data = json!({
+            "name": "test_function"
+        });
+
+        let expr = FilterExpr::StringOp {
+            field: "name".to_string(),
+            op: StringOp::Contains,
+            value: "func".to_string(),
+        };
+
+        assert!(evaluate(&expr, &data).unwrap());
+    }
+
+    #[test]
+    fn test_evaluate_nested_field() {
+        let data = json!({
+            "function": {
+                "name": "test",
+                "xrefs": {
+                    "count": 10
+                }
+            }
+        });
+
+        let expr = FilterExpr::Compare {
+            field: "function.xrefs.count".to_string(),
+            op: CompareOp::Greater,
+            value: Value::Integer(5),
+        };
+
+        assert!(evaluate(&expr, &data).unwrap());
+    }
+}
@@ -0,0 +1,137 @@
+pub mod parser;
+pub mod evaluator;
+
+use serde::{Deserialize, Serialize};
+use crate::error::Result;
+
+#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
+pub enum FilterExpr {
+    Compare {
+        field: String,
+        op: CompareOp,
+        value: Value,
+    },
+    StringOp {
+        field: String,
+        op: StringOp,
+        value: String,
+    },
+    Logical {
+        op: LogicalOp,
+        exprs: Vec<FilterExpr>,
+    },
+    Not(Box<FilterExpr>),
+    Exists {
+        field: String,
+        check: ExistenceCheck,
+    },
+    In {
+        field: String,
+        values: Vec<Value>,
+    },
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
+pub enum CompareOp {
+    Equal,
+    NotEqual,
+    Greater,
+    GreaterEqual,
+    Less,
+    LessEqual,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
+pub enum StringOp {
+    Contains,      // ~
+    StartsWith,    // ^
+    EndsWith,      // $
+    Regex,         // =~
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
+pub enum LogicalOp {
+    And,
+    Or,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
+pub enum ExistenceCheck {
+    Exists,
+    Empty,
+    Null,
+}
+
+#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
+pub enum Value {
+    String(String),
+    Number(f64),
+    Integer(i64),
+    Boolean(bool),
+    Hex(u64),
+}
+
+impl Value {
+    pub fn as_str(&self) -> Option<&str> {
+        match self {
+            Value::String(s) => Some(s),
+            _ => None,
+        }
+    }
+
+    pub fn as_f64(&self) -> Option<f64> {
+        match self {
+            Value::Number(n) => Some(*n),
+            Value::Integer(i) => Some(*i as f64),
+            Value::Hex(h) => Some(*h as f64),
+            _ => None,
+        }
+    }
+
+    pub fn as_i64(&self) -> Option<i64> {
+        match self {
+            Value::Integer(i) => Some(*i),
+            Value::Number(n) => Some(*n as i64),
+            Value::Hex(h) => Some(*h as i64),
+            _ => None,
+        }
+    }
+
+    pub fn as_bool(&self) -> Option<bool> {
+        match self {
+            Value::Boolean(b) => Some(*b),
+            _ => None,
+        }
+    }
+}
+
+pub struct Filter {
+    pub expr: FilterExpr,
+}
+
+impl Filter {
+    pub fn parse(input: &str) -> Result<Self> {
+        parser::parse_filter(input)
+    }
+
+    pub fn evaluate(&self, data: &serde_json::Value) -> Result<bool> {
+        evaluator::evaluate(&self.expr, data)
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_parse_simple_filter() {
+        let filter = Filter::parse("name=test").unwrap();
+        assert!(matches!(filter.expr, FilterExpr::Compare { .. }));
+    }
+
+    #[test]
+    fn test_parse_logical_filter() {
+        let filter = Filter::parse("name=test AND size>100").unwrap();
+        assert!(matches!(filter.expr, FilterExpr::Logical { .. }));
+    }
+}
@@ -0,0 +1,270 @@
+use pest::Parser;
+use pest_derive::Parser;
+use crate::error::{GhidraError, Result};
+use super::{Filter, FilterExpr, CompareOp, StringOp, LogicalOp, ExistenceCheck, Value};
+
+#[derive(Parser)]
+#[grammar = "filter.pest"]
+struct FilterParser;
+
+pub fn parse_filter(input: &str) -> Result<Filter> {
+    let pairs = FilterParser::parse(Rule::expr, input)
+        .map_err(|e| GhidraError::FilterParseError(format!("{}", e)))?;
+
+    let mut expr = None;
+    for pair in pairs {
+        if pair.as_rule() == Rule::expr {
+            for inner in pair.into_inner() {
+                if inner.as_rule() == Rule::logical_expr {
+                    expr = Some(parse_logical_expr(inner)?);
+                }
+            }
+        }
+    }
+
+    expr.map(|e| Filter { expr: e })
+        .ok_or_else(|| GhidraError::FilterParseError("Empty expression".to_string()))
+}
+
+fn parse_logical_expr(pair: pest::iterators::Pair<Rule>) -> Result<FilterExpr> {
+    let mut terms = Vec::new();
+    let mut ops = Vec::new();
+
+    for inner in pair.into_inner() {
+        match inner.as_rule() {
+            Rule::logical_term => {
+                terms.push(parse_logical_term(inner)?);
+            }
+            Rule::logical_op => {
+                let op_str = inner.as_str();
+                let op = match op_str {
+                    "AND" | "&&" => LogicalOp::And,
+                    "OR" | "||" => LogicalOp::Or,
+                    _ => return Err(GhidraError::FilterParseError(format!("Unknown operator: {}", op_str))),
+                };
+                ops.push(op);
+            }
+            _ => {}
+        }
+    }
+
+    if terms.is_empty() {
+        return Err(GhidraError::FilterParseError("No terms in logical expression".to_string()));
+    }
+
+    if terms.len() == 1 {
+        return Ok(terms.into_iter().next().unwrap());
+    }
+
+    // Build expression tree respecting precedence (AND before OR)
+    // For simplicity, we'll evaluate left-to-right for now
+    // TODO: Proper precedence handling
+    let mut result = terms[0].clone();
+    for (i, op) in ops.iter().enumerate() {
+        result = FilterExpr::Logical {
+            op: *op,
+            exprs: vec![result, terms[i + 1].clone()],
+        };
+    }
+
+    Ok(result)
+}
+
+fn parse_logical_term(pair: pest::iterators::Pair<Rule>) -> Result<FilterExpr> {
+    for inner in pair.into_inner() {
+        match inner.as_rule() {
+            Rule::logical_not => {
+                return parse_logical_not(inner);
+            }
+            Rule::logical_expr => {
+                return parse_logical_expr(inner);
+            }
+            Rule::comparison => {
+                return parse_comparison(inner);
+            }
+            _ => {}
+        }
+    }
+    Err(GhidraError::FilterParseError("Invalid logical term".to_string()))
+}
+
+fn parse_logical_not(pair: pest::iterators::Pair<Rule>) -> Result<FilterExpr> {
+    for inner in pair.into_inner() {
+        if inner.as_rule() == Rule::logical_term {
+            let term = parse_logical_term(inner)?;
+            return Ok(FilterExpr::Not(Box::new(term)));
+        }
+    }
+    Err(GhidraError::FilterParseError("Invalid NOT expression".to_string()))
+}
+
+fn parse_comparison(pair: pest::iterators::Pair<Rule>) -> Result<FilterExpr> {
+    let mut field = None;
+    let mut op = None;
+    let mut value = None;
+    let mut string_op = None;
+    let mut existence = None;
+    let mut in_values = None;
+
+    for inner in pair.into_inner() {
+        match inner.as_rule() {
+            Rule::field => {
+                field = Some(inner.as_str().to_string());
+            }
+            Rule::compare_op => {
+                let op_str = inner.as_str();
+                op = Some(match op_str {
+                    "=" => CompareOp::Equal,
+                    "!=" => CompareOp::NotEqual,
+                    ">" => CompareOp::Greater,
+                    ">=" => CompareOp::GreaterEqual,
+                    "<" => CompareOp::Less,
+                    "<=" => CompareOp::LessEqual,
+                    _ => return Err(GhidraError::FilterParseError(format!("Unknown compare op: {}", op_str))),
+                });
+            }
+            Rule::string_op => {
+                let op_str = inner.as_str();
+                string_op = Some(match op_str {
+                    "~" => StringOp::Contains,
+                    "^" => StringOp::StartsWith,
+                    "$" => StringOp::EndsWith,
+                    "=~" => StringOp::Regex,
+                    _ => return Err(GhidraError::FilterParseError(format!("Unknown string op: {}", op_str))),
+                });
+            }
+            Rule::value => {
+                value = Some(parse_value(inner)?);
+            }
+            Rule::string_value => {
+                value = Some(parse_value(inner)?);
+            }
+            Rule::existence_check => {
+                let check_str = inner.as_str();
+                existence = Some(match check_str {
+                    "EXISTS" => ExistenceCheck::Exists,
+                    "EMPTY" => ExistenceCheck::Empty,
+                    "NULL" => ExistenceCheck::Null,
+                    _ => return Err(GhidraError::FilterParseError(format!("Unknown existence check: {}", check_str))),
+                });
+            }
+            Rule::in_check => {
+                // Already handled field
+                continue;
+            }
+            Rule::value_list => {
+                let mut values = Vec::new();
+                for val_pair in inner.into_inner() {
+                    if val_pair.as_rule() == Rule::value {
+                        values.push(parse_value(val_pair)?);
+                    }
+                }
+                in_values = Some(values);
+            }
+            _ => {}
+        }
+    }
+
+    let field = field.ok_or_else(|| GhidraError::FilterParseError("Missing field".to_string()))?;
+
+    if let Some(existence_check) = existence {
+        return Ok(FilterExpr::Exists {
+            field,
+            check: existence_check,
+        });
+    }
+
+    if let Some(values) = in_values {
+        return Ok(FilterExpr::In { field, values });
+    }
+
+    if let Some(str_op) = string_op {
+        let val = value.ok_or_else(|| GhidraError::FilterParseError("Missing value".to_string()))?;
+        let val_str = match val {
+            Value::String(s) => s,
+            _ => return Err(GhidraError::FilterParseError("String operation requires string value".to_string())),
+        };
+        return Ok(FilterExpr::StringOp {
+            field,
+            op: str_op,
+            value: val_str,
+        });
+    }
+
+    if let Some(cmp_op) = op {
+        let val = value.ok_or_else(|| GhidraError::FilterParseError("Missing value".to_string()))?;
+        return Ok(FilterExpr::Compare {
+            field,
+            op: cmp_op,
+            value: val,
+        });
+    }
+
+    Err(GhidraError::FilterParseError("Invalid comparison".to_string()))
+}
+
+fn parse_value(pair: pest::iterators::Pair<Rule>) -> Result<Value> {
+    for inner in pair.into_inner() {
+        match inner.as_rule() {
+            Rule::number => {
+                let num_str = inner.as_str();
+                if num_str.contains('.') {
+                    let num = num_str.parse::<f64>()
+                        .map_err(|_| GhidraError::FilterParseError(format!("Invalid number: {}", num_str)))?;
+                    return Ok(Value::Number(num));
+                } else {
+                    let num = num_str.parse::<i64>()
+                        .map_err(|_| GhidraError::FilterParseError(format!("Invalid integer: {}", num_str)))?;
+                    return Ok(Value::Integer(num));
+                }
+            }
+            Rule::hex_number => {
+                let hex_str = inner.as_str().trim_start_matches("0x");
+                let num = u64::from_str_radix(hex_str, 16)
+                    .map_err(|_| GhidraError::FilterParseError(format!("Invalid hex number: {}", inner.as_str())))?;
+                return Ok(Value::Hex(num));
+            }
+            Rule::boolean => {
+                let bool_str = inner.as_str().to_lowercase();
+                return Ok(Value::Boolean(bool_str == "true"));
+            }
+            Rule::quoted_string => {
+                let s = inner.as_str();
+                let s = s.trim_matches(|c| c == '"' || c == '\'');
+                return Ok(Value::String(s.to_string()));
+            }
+            Rule::identifier => {
+                return Ok(Value::String(inner.as_str().to_string()));
+            }
+            _ => {}
+        }
+    }
+    Err(GhidraError::FilterParseError("Invalid value".to_string()))
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_parse_simple() {
+        let filter = parse_filter("name=test").unwrap();
+        assert!(matches!(filter.expr, FilterExpr::Compare { .. }));
+    }
+
+    #[test]
+    fn test_parse_and() {
+        let filter = parse_filter("name=test AND size>100").unwrap();
+        assert!(matches!(filter.expr, FilterExpr::Logical { .. }));
+    }
+
+    #[test]
+    fn test_parse_hex() {
+        let filter = parse_filter("address=0x401000").unwrap();
+        if let FilterExpr::Compare { value, .. } = filter.expr {
+            assert!(matches!(value, Value::Hex(0x401000)));
+        } else {
+            panic!("Expected Compare expression");
+        }
+    }
+}
@@ -0,0 +1,264 @@
+use serde::Serialize;
+use serde_json::Value as JsonValue;
+use crate::error::{GhidraError, Result};
+use comfy_table::{Table, presets::UTF8_FULL};
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum OutputFormat {
+    Full,
+    Compact,
+    Minimal,
+    Json,
+    JsonCompact,
+    JsonStream,
+    Csv,
+    Tsv,
+    Table,
+    Ids,
+    Count,
+    Tree,
+    Hex,
+    Asm,
+    C,
+}
+
+impl OutputFormat {
+    pub fn from_str(s: &str) -> Result<Self> {
+        match s.to_lowercase().as_str() {
+            "full" => Ok(Self::Full),
+            "compact" => Ok(Self::Compact),
+            "minimal" => Ok(Self::Minimal),
+            "json" => Ok(Self::Json),
+            "json-compact" => Ok(Self::JsonCompact),
+            "json-stream" | "ndjson" => Ok(Self::JsonStream),
+            "csv" => Ok(Self::Csv),
+            "tsv" => Ok(Self::Tsv),
+            "table" => Ok(Self::Table),
+            "ids" => Ok(Self::Ids),
+            "count" => Ok(Self::Count),
+            "tree" => Ok(Self::Tree),
+            "hex" => Ok(Self::Hex),
+            "asm" => Ok(Self::Asm),
+            "c" => Ok(Self::C),
+            _ => Err(GhidraError::InvalidFormat(format!("Unknown format: {}", s))),
+        }
+    }
+
+    pub fn is_human_friendly(&self) -> bool {
+        matches!(self, Self::Full | Self::Compact | Self::Table | Self::Tree)
+    }
+
+    pub fn is_machine_friendly(&self) -> bool {
+        matches!(self, Self::Json | Self::JsonCompact | Self::JsonStream | Self::Csv | Self::Tsv)
+    }
+}
+
+pub trait Formatter {
+    fn format<T: Serialize>(&self, data: &[T], format: OutputFormat) -> Result<String>;
+}
+
+pub struct DefaultFormatter;
+
+impl Formatter for DefaultFormatter {
+    fn format<T: Serialize>(&self, data: &[T], format: OutputFormat) -> Result<String> {
+        match format {
+            OutputFormat::Json => {
+                serde_json::to_string_pretty(data).map_err(|e| e.into())
+            }
+            OutputFormat::JsonCompact => {
+                serde_json::to_string(data).map_err(|e| e.into())
+            }
+            OutputFormat::JsonStream => {
+                let mut result = String::new();
+                for item in data {
+                    let json = serde_json::to_string(item)?;
+                    result.push_str(&json);
+                    result.push('\n');
+                }
+                Ok(result)
+            }
+            OutputFormat::Count => {
+                Ok(format!("{}", data.len()))
+            }
+            OutputFormat::Table => {
+                format_table(data)
+            }
+            OutputFormat::Csv => {
+                format_csv(data, ',')
+            }
+            OutputFormat::Tsv => {
+                format_csv(data, '\t')
+            }
+            OutputFormat::Minimal | OutputFormat::Ids => {
+                format_minimal(data)
+            }
+            _ => {
+                // For other formats, default to JSON
+                serde_json::to_string_pretty(data).map_err(|e| e.into())
+            }
+        }
+    }
+}
+
+fn format_table<T: Serialize>(data: &[T]) -> Result<String> {
+    if data.is_empty() {
+        return Ok("No results".to_string());
+    }
+
+    // Convert to JSON values to inspect structure
+    let json_data: Vec<JsonValue> = data.iter()
+        .map(|item| serde_json::to_value(item))
+        .collect::<std::result::Result<Vec<_>, _>>()?;
+
+    if json_data.is_empty() {
+        return Ok("No results".to_string());
+    }
+
+    // Get all keys from first object
+    let keys = if let Some(JsonValue::Object(map)) = json_data.first() {
+        map.keys().cloned().collect::<Vec<_>>()
+    } else {
+        return Ok(format!("{} results", data.len()));
+    };
+
+    let mut table = Table::new();
+    table.load_preset(UTF8_FULL);
+
+    // Add header
+    table.set_header(&keys);
+
+    // Add rows
+    for item in &json_data {
+        if let JsonValue::Object(map) = item {
+            let row: Vec<String> = keys.iter()
+                .map(|k| {
+                    map.get(k)
+                        .map(|v| format_json_value(v))
+                        .unwrap_or_else(|| "".to_string())
+                })
+                .collect();
+            table.add_row(row);
+        }
+    }
+
+    Ok(table.to_string())
+}
+
+fn format_csv<T: Serialize>(data: &[T], delimiter: char) -> Result<String> {
+    if data.is_empty() {
+        return Ok(String::new());
+    }
+
+    let json_data: Vec<JsonValue> = data.iter()
+        .map(|item| serde_json::to_value(item))
+        .collect::<std::result::Result<Vec<_>, _>>()?;
+
+    if json_data.is_empty() {
+        return Ok(String::new());
+    }
+
+    let keys = if let Some(JsonValue::Object(map)) = json_data.first() {
+        map.keys().cloned().collect::<Vec<_>>()
+    } else {
+        return Ok(String::new());
+    };
+
+    let mut result = String::new();
+
+    // Header
+    result.push_str(&keys.join(&delimiter.to_string()));
+    result.push('\n');
+
+    // Rows
+    for item in &json_data {
+        if let JsonValue::Object(map) = item {
+            let row: Vec<String> = keys.iter()
+                .map(|k| {
+                    map.get(k)
+                        .map(|v| format_json_value(v))
+                        .unwrap_or_else(|| "".to_string())
+                })
+                .collect();
+            result.push_str(&row.join(&delimiter.to_string()));
+            result.push('\n');
+        }
+    }
+
+    Ok(result)
+}
+
+fn format_minimal<T: Serialize>(data: &[T]) -> Result<String> {
+    let json_data: Vec<JsonValue> = data.iter()
+        .map(|item| serde_json::to_value(item))
+        .collect::<std::result::Result<Vec<_>, _>>()?;
+
+    let mut result = String::new();
+
+    for item in &json_data {
+        if let JsonValue::Object(map) = item {
+            // Try to get address or name or first field
+            let value = map.get("address")
+                .or_else(|| map.get("name"))
+                .or_else(|| map.get("id"))
+                .or_else(|| map.values().next())
+                .map(|v| format_json_value(v))
+                .unwrap_or_else(|| "".to_string());
+
+            result.push_str(&value);
+            result.push('\n');
+        } else {
+            result.push_str(&format_json_value(item));
+            result.push('\n');
+        }
+    }
+
+    Ok(result)
+}
+
+fn format_json_value(value: &JsonValue) -> String {
+    match value {
+        JsonValue::Null => "null".to_string(),
+        JsonValue::Bool(b) => b.to_string(),
+        JsonValue::Number(n) => n.to_string(),
+        JsonValue::String(s) => s.clone(),
+        JsonValue::Array(arr) => {
+            format!("[{}]", arr.iter().map(format_json_value).collect::<Vec<_>>().join(", "))
+        }
+        JsonValue::Object(_) => serde_json::to_string(value).unwrap_or_else(|_| "{}".to_string()),
+    }
+}
+
+pub fn auto_detect_format(is_tty: bool) -> OutputFormat {
+    if is_tty {
+        OutputFormat::Table
+    } else {
+        OutputFormat::JsonCompact
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use serde_json::json;
+
+    #[test]
+    fn test_format_json() {
+        let data = vec![
+            json!({"name": "test", "value": 123}),
+        ];
+        let formatter = DefaultFormatter;
+        let result = formatter.format(&data, OutputFormat::Json).unwrap();
+        assert!(result.contains("test"));
+    }
+
+    #[test]
+    fn test_format_count() {
+        let data = vec![
+            json!({"name": "test1"}),
+            json!({"name": "test2"}),
+        ];
+        let formatter = DefaultFormatter;
+        let result = formatter.format(&data, OutputFormat::Count).unwrap();
+        assert_eq!(result, "2");
+    }
+}
@@ -0,0 +1,168 @@
+use serde::{Deserialize, Serialize};
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct Function {
+    pub name: String,
+    pub address: String,
+    pub size: u64,
+    pub signature: Option<String>,
+    pub entry_point: String,
+    pub calling_convention: Option<String>,
+    #[serde(default)]
+    pub parameters: Vec<Parameter>,
+    #[serde(default)]
+    pub local_variables: Vec<LocalVariable>,
+    #[serde(default)]
+    pub calls: Vec<String>,
+    #[serde(default)]
+    pub called_by: Vec<String>,
+    pub decompiled: Option<String>,
+    pub comment: Option<String>,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct Parameter {
+    pub name: String,
+    pub data_type: String,
+    pub ordinal: u32,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct LocalVariable {
+    pub name: String,
+    pub data_type: String,
+    pub stack_offset: Option<i32>,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct StringData {
+    pub address: String,
+    pub value: String,
+    pub length: usize,
+    pub encoding: String,
+    #[serde(default)]
+    pub references: Vec<String>,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct Symbol {
+    pub name: String,
+    pub address: String,
+    pub symbol_type: String,
+    pub namespace: Option<String>,
+    pub source: String,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct Import {
+    pub name: String,
+    pub address: String,
+    pub library: String,
+    pub ordinal: Option<u32>,
+    pub is_external: bool,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct Export {
+    pub name: String,
+    pub address: String,
+    pub ordinal: Option<u32>,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct XRef {
+    pub from: String,
+    pub to: String,
+    pub ref_type: String,
+    pub from_function: Option<String>,
+    pub to_function: Option<String>,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct MemoryBlock {
+    pub name: String,
+    pub start: String,
+    pub end: String,
+    pub size: u64,
+    pub permissions: String,
+    pub is_initialized: bool,
+    pub is_loaded: bool,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct Section {
+    pub name: String,
+    pub address: String,
+    pub size: u64,
+    pub virtual_address: String,
+    pub file_offset: u64,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct Comment {
+    pub address: String,
+    pub comment_type: String,
+    pub text: String,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct DataType {
+    pub name: String,
+    pub category: String,
+    pub size: Option<u64>,
+    pub description: Option<String>,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct Instruction {
+    pub address: String,
+    pub mnemonic: String,
+    pub operands: String,
+    pub bytes: String,
+    pub length: u32,
+    pub flow_type: String,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct BasicBlock {
+    pub start: String,
+    pub end: String,
+    pub size: u64,
+    pub instruction_count: u32,
+    #[serde(default)]
+    pub successors: Vec<String>,
+    #[serde(default)]
+    pub predecessors: Vec<String>,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct ProgramInfo {
+    pub name: String,
+    pub executable_path: String,
+    pub executable_format: String,
+    pub compiler: Option<String>,
+    pub language: String,
+    pub creation_date: Option<String>,
+    pub image_base: String,
+    pub min_address: String,
+    pub max_address: String,
+    pub function_count: usize,
+    pub instruction_count: usize,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+#[serde(tag = "type")]
+pub enum GhidraData {
+    Function(Function),
+    String(StringData),
+    Symbol(Symbol),
+    Import(Import),
+    Export(Export),
+    XRef(XRef),
+    MemoryBlock(MemoryBlock),
+    Section(Section),
+    Comment(Comment),
+    DataType(DataType),
+    Instruction(Instruction),
+    BasicBlock(BasicBlock),
+}
@@ -0,0 +1,211 @@
+use std::path::Path;
+use std::process::Command;
+use serde_json::Value as JsonValue;
+use crate::error::{GhidraError, Result};
+use super::GhidraClient;
+use super::scripts;
+
+pub struct HeadlessExecutor<'a> {
+    client: &'a GhidraClient,
+}
+
+impl<'a> HeadlessExecutor<'a> {
+    pub fn new(client: &'a GhidraClient) -> Self {
+        Self { client }
+    }
+
+    pub fn execute_script(
+        &self,
+        project_name: &str,
+        program_name: &str,
+        script_content: &str,
+        script_name: &str,
+        args: &[String],
+    ) -> Result<JsonValue> {
+        // Save script to disk
+        let scripts_dir = self.get_scripts_dir()?;
+        let script_path = scripts::save_script(script_name, script_content, &scripts_dir)?;
+
+        // Execute script
+        let output = self.run_ghidra_script(project_name, program_name, &script_path, args)?;
+
+        // Parse JSON output
+        let json: JsonValue = serde_json::from_str(&output)
+            .map_err(|e| GhidraError::ExecutionFailed(format!("Failed to parse script output: {}", e)))?;
+
+        Ok(json)
+    }
+
+    fn run_ghidra_script(
+        &self,
+        project_name: &str,
+        program_name: &str,
+        script_path: &Path,
+        args: &[String],
+    ) -> Result<String> {
+        let project_path = self.client.get_project_path(project_name);
+        let headless = self.client.get_headless_script();
+
+        let mut cmd = Command::new(&headless);
+        cmd.arg(project_path.to_str().unwrap())
+            .arg(project_name)
+            .arg("-process")
+            .arg(program_name)
+            .arg("-noanalysis")
+            .arg("-scriptPath")
+            .arg(script_path.parent().unwrap().to_str().unwrap())
+            .arg("-postScript")
+            .arg(script_path.file_name().unwrap().to_str().unwrap());
+
+        for arg in args {
+            cmd.arg(arg);
+        }
+
+        // Capture output
+        let output = cmd.output()?;
+
+        if !output.status.success() {
+            let stderr = String::from_utf8_lossy(&output.stderr);
+            return Err(GhidraError::ExecutionFailed(format!("Script execution failed: {}", stderr)));
+        }
+
+        // Extract JSON from output (Ghidra adds some logging we need to skip)
+        let stdout = String::from_utf8_lossy(&output.stdout);
+        let json_output = self.extract_json_from_output(&stdout)?;
+
+        Ok(json_output)
+    }
+
+    fn extract_json_from_output(&self, output: &str) -> Result<String> {
+        // Find the JSON output in the Ghidra output
+        // Look for lines starting with { or [
+        let lines: Vec<&str> = output.lines().collect();
+
+        let mut json_start = None;
+        let mut json_end = None;
+        let mut brace_count = 0;
+
+        for (i, line) in lines.iter().enumerate() {
+            let trimmed = line.trim();
+
+            if json_start.is_none() && (trimmed.starts_with('{') || trimmed.starts_with('[')) {
+                json_start = Some(i);
+                brace_count = trimmed.chars().filter(|&c| c == '{' || c == '[').count() as i32;
+                brace_count -= trimmed.chars().filter(|&c| c == '}' || c == ']').count() as i32;
+
+                if brace_count == 0 {
+                    json_end = Some(i);
+                    break;
+                }
+            } else if json_start.is_some() {
+                brace_count += trimmed.chars().filter(|&c| c == '{' || c == '[').count() as i32;
+                brace_count -= trimmed.chars().filter(|&c| c == '}' || c == ']').count() as i32;
+
+                if brace_count == 0 {
+                    json_end = Some(i);
+                    break;
+                }
+            }
+        }
+
+        if let (Some(start), Some(end)) = (json_start, json_end) {
+            let json_lines = &lines[start..=end];
+            Ok(json_lines.join("\n"))
+        } else {
+            Err(GhidraError::ExecutionFailed("Could not find JSON in script output".to_string()))
+        }
+    }
+
+    fn get_scripts_dir(&self) -> Result<std::path::PathBuf> {
+        let config_dir = dirs::config_dir()
+            .ok_or_else(|| GhidraError::ConfigError("Could not determine config directory".to_string()))?;
+
+        let scripts_dir = config_dir.join("ghidra-cli").join("scripts");
+
+        if !scripts_dir.exists() {
+            std::fs::create_dir_all(&scripts_dir)?;
+        }
+
+        Ok(scripts_dir)
+    }
+
+    pub fn list_functions(&self, project_name: &str, program_name: &str) -> Result<JsonValue> {
+        self.execute_script(
+            project_name,
+            program_name,
+            scripts::get_list_functions_script(),
+            "list_functions",
+            &[],
+        )
+    }
+
+    pub fn decompile_function(&self, project_name: &str, program_name: &str, address: &str) -> Result<JsonValue> {
+        self.execute_script(
+            project_name,
+            program_name,
+            scripts::get_decompile_function_script(),
+            "decompile_function",
+            &[address.to_string()],
+        )
+    }
+
+    pub fn list_strings(&self, project_name: &str, program_name: &str) -> Result<JsonValue> {
+        self.execute_script(
+            project_name,
+            program_name,
+            scripts::get_list_strings_script(),
+            "list_strings",
+            &[],
+        )
+    }
+
+    pub fn list_imports(&self, project_name: &str, program_name: &str) -> Result<JsonValue> {
+        self.execute_script(
+            project_name,
+            program_name,
+            scripts::get_list_imports_script(),
+            "list_imports",
+            &[],
+        )
+    }
+
+    pub fn list_exports(&self, project_name: &str, program_name: &str) -> Result<JsonValue> {
+        self.execute_script(
+            project_name,
+            program_name,
+            scripts::get_list_exports_script(),
+            "list_exports",
+            &[],
+        )
+    }
+
+    pub fn get_memory_map(&self, project_name: &str, program_name: &str) -> Result<JsonValue> {
+        self.execute_script(
+            project_name,
+            program_name,
+            scripts::get_memory_map_script(),
+            "memory_map",
+            &[],
+        )
+    }
+
+    pub fn get_program_info(&self, project_name: &str, program_name: &str) -> Result<JsonValue> {
+        self.execute_script(
+            project_name,
+            program_name,
+            scripts::get_program_info_script(),
+            "program_info",
+            &[],
+        )
+    }
+
+    pub fn get_xrefs_to(&self, project_name: &str, program_name: &str, address: &str) -> Result<JsonValue> {
+        self.execute_script(
+            project_name,
+            program_name,
+            scripts::get_xrefs_to_script(),
+            "xrefs_to",
+            &[address.to_string()],
+        )
+    }
+}
@@ -0,0 +1,212 @@
+pub mod headless;
+pub mod data;
+pub mod scripts;
+
+use std::path::{Path, PathBuf};
+use std::process::Command;
+use crate::config::Config;
+use crate::error::{GhidraError, Result};
+
+pub struct GhidraClient {
+    config: Config,
+    install_dir: PathBuf,
+    project_dir: PathBuf,
+}
+
+impl GhidraClient {
+    pub fn new(config: Config) -> Result<Self> {
+        let install_dir = config.get_ghidra_install_dir()?;
+        let project_dir = config.get_project_dir()?;
+
+        // Create project directory if it doesn't exist
+        if !project_dir.exists() {
+            std::fs::create_dir_all(&project_dir)?;
+        }
+
+        Ok(Self {
+            config,
+            install_dir,
+            project_dir,
+        })
+    }
+
+    pub fn get_headless_script(&self) -> PathBuf {
+        let support_dir = self.install_dir.join("support");
+
+        #[cfg(target_os = "windows")]
+        {
+            support_dir.join("analyzeHeadless.bat")
+        }
+
+        #[cfg(not(target_os = "windows"))]
+        {
+            support_dir.join("analyzeHeadless")
+        }
+    }
+
+    pub fn verify_installation(&self) -> Result<()> {
+        let headless = self.get_headless_script();
+
+        if !headless.exists() {
+            return Err(GhidraError::GhidraNotFound);
+        }
+
+        Ok(())
+    }
+
+    pub fn get_project_path(&self, project_name: &str) -> PathBuf {
+        self.project_dir.join(project_name)
+    }
+
+    pub fn project_exists(&self, project_name: &str) -> bool {
+        let project_path = self.get_project_path(project_name);
+        project_path.exists() && project_path.join(format!("{}.rep", project_name)).exists()
+    }
+
+    pub fn create_project(&self, project_name: &str) -> Result<()> {
+        let project_path = self.get_project_path(project_name);
+
+        if self.project_exists(project_name) {
+            return Ok(());
+        }
+
+        std::fs::create_dir_all(&project_path)?;
+
+        // Run headless to create project
+        let headless = self.get_headless_script();
+        let output = Command::new(&headless)
+            .arg(project_path.to_str().unwrap())
+            .arg(project_name)
+            .output()?;
+
+        if !output.status.success() {
+            return Err(GhidraError::ExecutionFailed(
+                String::from_utf8_lossy(&output.stderr).to_string()
+            ));
+        }
+
+        Ok(())
+    }
+
+    pub fn import_binary(&self, project_name: &str, binary_path: &Path, program_name: Option<&str>) -> Result<String> {
+        if !self.project_exists(project_name) {
+            self.create_project(project_name)?;
+        }
+
+        let program_name = program_name.unwrap_or_else(|| {
+            binary_path.file_name()
+                .and_then(|n| n.to_str())
+                .unwrap_or("program")
+        });
+
+        let project_path = self.get_project_path(project_name);
+        let headless = self.get_headless_script();
+
+        let output = Command::new(&headless)
+            .arg(project_path.to_str().unwrap())
+            .arg(project_name)
+            .arg("-import")
+            .arg(binary_path.to_str().unwrap())
+            .arg("-overwrite")
+            .output()?;
+
+        if !output.status.success() {
+            return Err(GhidraError::ExecutionFailed(
+                String::from_utf8_lossy(&output.stderr).to_string()
+            ));
+        }
+
+        Ok(program_name.to_string())
+    }
+
+    pub fn analyze_program(&self, project_name: &str, program_name: &str) -> Result<()> {
+        let project_path = self.get_project_path(project_name);
+        let headless = self.get_headless_script();
+
+        let output = Command::new(&headless)
+            .arg(project_path.to_str().unwrap())
+            .arg(project_name)
+            .arg("-process")
+            .arg(program_name)
+            .arg("-noanalysis")  // Don't auto-analyze
+            .arg("-scriptPath")
+            .arg(self.get_scripts_dir()?.to_str().unwrap())
+            .output()?;
+
+        if !output.status.success() {
+            return Err(GhidraError::ExecutionFailed(
+                String::from_utf8_lossy(&output.stderr).to_string()
+            ));
+        }
+
+        Ok(())
+    }
+
+    pub fn run_script(&self, project_name: &str, program_name: &str, script_path: &Path, args: &[String]) -> Result<String> {
+        let project_path = self.get_project_path(project_name);
+        let headless = self.get_headless_script();
+
+        let mut cmd = Command::new(&headless);
+        cmd.arg(project_path.to_str().unwrap())
+            .arg(project_name)
+            .arg("-process")
+            .arg(program_name)
+            .arg("-scriptPath")
+            .arg(script_path.parent().unwrap().to_str().unwrap())
+            .arg("-postScript")
+            .arg(script_path.file_name().unwrap().to_str().unwrap());
+
+        for arg in args {
+            cmd.arg(arg);
+        }
+
+        let output = cmd.output()?;
+
+        if !output.status.success() {
+            return Err(GhidraError::ExecutionFailed(
+                String::from_utf8_lossy(&output.stderr).to_string()
+            ));
+        }
+
+        Ok(String::from_utf8_lossy(&output.stdout).to_string())
+    }
+
+    fn get_scripts_dir(&self) -> Result<PathBuf> {
+        let config_dir = dirs::config_dir()
+            .ok_or_else(|| GhidraError::ConfigError("Could not determine config directory".to_string()))?;
+
+        let scripts_dir = config_dir.join("ghidra-cli").join("scripts");
+
+        if !scripts_dir.exists() {
+            std::fs::create_dir_all(&scripts_dir)?;
+        }
+
+        Ok(scripts_dir)
+    }
+
+    pub fn get_install_dir(&self) -> &Path {
+        &self.install_dir
+    }
+
+    pub fn get_project_dir(&self) -> &Path {
+        &self.project_dir
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_ghidra_client_creation() {
+        // This will fail if GHIDRA_INSTALL_DIR is not set, which is expected
+        let config = Config::default();
+        let result = GhidraClient::new(config);
+
+        // We can't test this properly without a Ghidra installation
+        // Just verify the error is what we expect
+        if result.is_err() {
+            assert!(matches!(result.unwrap_err(), GhidraError::GhidraNotFound));
+        }
+    }
+}
@@ -0,0 +1,310 @@
+/// Built-in Ghidra scripts for data extraction
+/// These are Python scripts that will be written to disk and executed by Ghidra headless
+
+pub fn get_list_functions_script() -> &'static str {
+    r#"
+# List all functions in the program
+# @category Analysis
+
+import json
+
+functions = []
+function_manager = currentProgram.getFunctionManager()
+
+for func in function_manager.getFunctions(True):
+    entry = func.getEntryPoint()
+    body = func.getBody()
+
+    func_data = {
+        "name": func.getName(),
+        "address": entry.toString(),
+        "size": body.getNumAddresses(),
+        "entry_point": entry.toString(),
+        "signature": func.getPrototypeString(False, False) if func.getSignature() else None,
+        "calling_convention": func.getCallingConventionName(),
+        "comment": func.getComment()
+    }
+
+    # Get called functions
+    called = []
+    refs = func.getBody().getAddresses(True)
+    for addr in refs:
+        for ref in currentProgram.getReferenceManager().getReferencesFrom(addr):
+            if ref.getReferenceType().isCall():
+                to_addr = ref.getToAddress()
+                to_func = function_manager.getFunctionAt(to_addr)
+                if to_func:
+                    called.append(to_func.getName())
+
+    func_data["calls"] = list(set(called))
+
+    # Get callers
+    callers = []
+    refs_to = currentProgram.getReferenceManager().getReferencesTo(entry)
+    for ref in refs_to:
+        if ref.getReferenceType().isCall():
+            from_addr = ref.getFromAddress()
+            from_func = function_manager.getFunctionContaining(from_addr)
+            if from_func:
+                callers.append(from_func.getName())
+
+    func_data["called_by"] = list(set(callers))
+
+    functions.append(func_data)
+
+print(json.dumps(functions, indent=2))
+"#
+}
+
+pub fn get_decompile_function_script() -> &'static str {
+    r#"
+# Decompile a specific function
+# @category Analysis
+
+import json
+from ghidra.app.decompiler import DecompInterface
+from ghidra.util.task import ConsoleTaskMonitor
+
+# Get function address from args
+if len(args) < 1:
+    print(json.dumps({"error": "No address provided"}))
+    exit(1)
+
+addr_str = args[0]
+addr = currentProgram.getAddressFactory().getAddress(addr_str)
+
+function_manager = currentProgram.getFunctionManager()
+func = function_manager.getFunctionContaining(addr)
+
+if not func:
+    print(json.dumps({"error": "No function at address " + addr_str}))
+    exit(1)
+
+# Decompile
+decompiler = DecompInterface()
+decompiler.openProgram(currentProgram)
+
+monitor = ConsoleTaskMonitor()
+results = decompiler.decompileFunction(func, 30, monitor)
+
+if results.decompileCompleted():
+    code = results.getDecompiledFunction().getC()
+
+    result = {
+        "name": func.getName(),
+        "address": func.getEntryPoint().toString(),
+        "signature": func.getPrototypeString(False, False),
+        "code": code
+    }
+
+    print(json.dumps(result, indent=2))
+else:
+    print(json.dumps({"error": "Decompilation failed"}))
+"#
+}
+
+pub fn get_list_strings_script() -> &'static str {
+    r#"
+# List all strings in the program
+# @category Analysis
+
+import json
+
+strings = []
+string_table = currentProgram.getListing().getDefinedData(True)
+
+for data in string_table:
+    if data.hasStringValue():
+        string_data = {
+            "address": data.getAddress().toString(),
+            "value": data.getValue().toString(),
+            "length": len(data.getValue().toString()),
+            "encoding": "ascii"
+        }
+
+        # Get references to this string
+        refs = []
+        refs_to = currentProgram.getReferenceManager().getReferencesTo(data.getAddress())
+        for ref in refs_to:
+            refs.append(ref.getFromAddress().toString())
+
+        string_data["references"] = refs
+        strings.append(string_data)
+
+print(json.dumps(strings, indent=2))
+"#
+}
+
+pub fn get_list_imports_script() -> &'static str {
+    r#"
+# List all imports in the program
+# @category Analysis
+
+import json
+
+imports = []
+symbol_table = currentProgram.getSymbolTable()
+external_manager = currentProgram.getExternalManager()
+
+for symbol in symbol_table.getExternalSymbols():
+    external_location = external_manager.getExternalLocation(symbol)
+
+    if external_location:
+        import_data = {
+            "name": symbol.getName(),
+            "address": symbol.getAddress().toString(),
+            "library": external_location.getLibraryName(),
+            "is_external": True
+        }
+        imports.append(import_data)
+
+print(json.dumps(imports, indent=2))
+"#
+}
+
+pub fn get_list_exports_script() -> &'static str {
+    r#"
+# List all exports in the program
+# @category Analysis
+
+import json
+
+exports = []
+symbol_table = currentProgram.getSymbolTable()
+
+for symbol in symbol_table.getSymbolIterator():
+    if symbol.isExternalEntryPoint():
+        export_data = {
+            "name": symbol.getName(),
+            "address": symbol.getAddress().toString()
+        }
+        exports.append(export_data)
+
+print(json.dumps(exports, indent=2))
+"#
+}
+
+pub fn get_memory_map_script() -> &'static str {
+    r#"
+# Get memory map
+# @category Analysis
+
+import json
+
+blocks = []
+memory = currentProgram.getMemory()
+
+for block in memory.getBlocks():
+    block_data = {
+        "name": block.getName(),
+        "start": block.getStart().toString(),
+        "end": block.getEnd().toString(),
+        "size": block.getSize(),
+        "permissions": "",
+        "is_initialized": block.isInitialized(),
+        "is_loaded": block.isLoaded()
+    }
+
+    # Build permissions string
+    perms = ""
+    if block.isRead():
+        perms += "r"
+    if block.isWrite():
+        perms += "w"
+    if block.isExecute():
+        perms += "x"
+
+    block_data["permissions"] = perms
+    blocks.append(block_data)
+
+print(json.dumps(blocks, indent=2))
+"#
+}
+
+pub fn get_program_info_script() -> &'static str {
+    r#"
+# Get program information
+# @category Analysis
+
+import json
+
+info = {
+    "name": currentProgram.getName(),
+    "executable_path": currentProgram.getExecutablePath(),
+    "executable_format": currentProgram.getExecutableFormat(),
+    "compiler": currentProgram.getCompiler() if currentProgram.getCompiler() else None,
+    "language": currentProgram.getLanguage().toString(),
+    "image_base": currentProgram.getImageBase().toString(),
+    "min_address": currentProgram.getMinAddress().toString(),
+    "max_address": currentProgram.getMaxAddress().toString()
+}
+
+# Count functions and instructions
+function_manager = currentProgram.getFunctionManager()
+info["function_count"] = function_manager.getFunctionCount()
+
+instruction_count = 0
+listing = currentProgram.getListing()
+for instruction in listing.getInstructions(True):
+    instruction_count += 1
+
+info["instruction_count"] = instruction_count
+
+print(json.dumps(info, indent=2))
+"#
+}
+
+pub fn get_xrefs_to_script() -> &'static str {
+    r#"
+# Get cross-references to an address
+# @category Analysis
+
+import json
+
+if len(args) < 1:
+    print(json.dumps({"error": "No address provided"}))
+    exit(1)
+
+addr_str = args[0]
+addr = currentProgram.getAddressFactory().getAddress(addr_str)
+
+xrefs = []
+refs = currentProgram.getReferenceManager().getReferencesTo(addr)
+function_manager = currentProgram.getFunctionManager()
+
+for ref in refs:
+    from_addr = ref.getFromAddress()
+    from_func = function_manager.getFunctionContaining(from_addr)
+    to_func = function_manager.getFunctionContaining(addr)
+
+    xref_data = {
+        "from": from_addr.toString(),
+        "to": addr.toString(),
+        "ref_type": ref.getReferenceType().toString(),
+        "from_function": from_func.getName() if from_func else None,
+        "to_function": to_func.getName() if to_func else None
+    }
+    xrefs.append(xref_data)
+
+print(json.dumps(xrefs, indent=2))
+"#
+}
+
+/// Save a script to disk
+pub fn save_script(name: &str, content: &str, scripts_dir: &std::path::Path) -> crate::error::Result<std::path::PathBuf> {
+    let script_path = scripts_dir.join(format!("{}.py", name));
+    std::fs::write(&script_path, content)?;
+    Ok(script_path)
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_scripts_not_empty() {
+        assert!(!get_list_functions_script().is_empty());
+        assert!(!get_decompile_function_script().is_empty());
+        assert!(!get_list_strings_script().is_empty());
+    }
+}
@@ -0,0 +1,356 @@
+use serde_json::Value as JsonValue;
+use crate::error::{GhidraError, Result};
+use crate::filter::{Filter, FilterExpr};
+use crate::format::{OutputFormat, Formatter, DefaultFormatter};
+use crate::ghidra::GhidraClient;
+use crate::ghidra::headless::HeadlessExecutor;
+
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub enum DataType {
+    Functions,
+    Strings,
+    Symbols,
+    Imports,
+    Exports,
+    XRefs,
+    Memory,
+    Sections,
+    Comments,
+    Types,
+    Instructions,
+    BasicBlocks,
+    CallGraph,
+    Data,
+    References,
+}
+
+impl DataType {
+    pub fn from_str(s: &str) -> Result<Self> {
+        match s.to_lowercase().as_str() {
+            "functions" | "function" | "fn" => Ok(Self::Functions),
+            "strings" | "string" | "str" => Ok(Self::Strings),
+            "symbols" | "symbol" | "sym" => Ok(Self::Symbols),
+            "imports" | "import" => Ok(Self::Imports),
+            "exports" | "export" => Ok(Self::Exports),
+            "xrefs" | "xref" | "crossrefs" => Ok(Self::XRefs),
+            "memory" | "mem" => Ok(Self::Memory),
+            "sections" | "section" => Ok(Self::Sections),
+            "comments" | "comment" => Ok(Self::Comments),
+            "types" | "type" => Ok(Self::Types),
+            "instructions" | "instruction" | "insn" => Ok(Self::Instructions),
+            "basicblocks" | "basic-blocks" | "blocks" => Ok(Self::BasicBlocks),
+            "callgraph" | "call-graph" => Ok(Self::CallGraph),
+            "data" => Ok(Self::Data),
+            "references" | "refs" => Ok(Self::References),
+            _ => Err(GhidraError::InvalidDataType(format!("Unknown data type: {}", s))),
+        }
+    }
+}
+
+pub struct Query {
+    pub data_type: DataType,
+    pub filter: Option<Filter>,
+    pub fields: Option<FieldSelector>,
+    pub format: OutputFormat,
+    pub limit: Option<usize>,
+    pub offset: Option<usize>,
+    pub sort: Option<Vec<SortKey>>,
+    pub count_only: bool,
+}
+
+impl Query {
+    pub fn new(data_type: DataType) -> Self {
+        Self {
+            data_type,
+            filter: None,
+            fields: None,
+            format: OutputFormat::Json,
+            limit: None,
+            offset: None,
+            sort: None,
+            count_only: false,
+        }
+    }
+
+    pub fn with_filter(mut self, filter: Filter) -> Self {
+        self.filter = Some(filter);
+        self
+    }
+
+    pub fn with_format(mut self, format: OutputFormat) -> Self {
+        self.format = format;
+        self
+    }
+
+    pub fn with_limit(mut self, limit: usize) -> Self {
+        self.limit = Some(limit);
+        self
+    }
+
+    pub fn with_offset(mut self, offset: usize) -> Self {
+        self.offset = Some(offset);
+        self
+    }
+
+    pub fn count_only(mut self) -> Self {
+        self.count_only = true;
+        self
+    }
+
+    pub fn execute(&self, client: &GhidraClient, project: &str, program: &str) -> Result<String> {
+        let executor = HeadlessExecutor::new(client);
+
+        // Fetch data based on type
+        let data = match self.data_type {
+            DataType::Functions => executor.list_functions(project, program)?,
+            DataType::Strings => executor.list_strings(project, program)?,
+            DataType::Imports => executor.list_imports(project, program)?,
+            DataType::Exports => executor.list_exports(project, program)?,
+            DataType::Memory => executor.get_memory_map(project, program)?,
+            _ => {
+                return Err(GhidraError::Other(format!(
+                    "Data type {:?} not yet implemented",
+                    self.data_type
+                )));
+            }
+        };
+
+        // Data should be an array
+        let data_array = match data {
+            JsonValue::Array(arr) => arr,
+            _ => {
+                return Err(GhidraError::ExecutionFailed(
+                    "Expected array from Ghidra script".to_string()
+                ));
+            }
+        };
+
+        // Apply filter
+        let filtered = if let Some(filter) = &self.filter {
+            self.apply_filter(&data_array, filter)?
+        } else {
+            data_array
+        };
+
+        // Apply field selection
+        let selected = if let Some(fields) = &self.fields {
+            self.select_fields(&filtered, fields)?
+        } else {
+            filtered
+        };
+
+        // Apply sorting
+        let sorted = if let Some(sort) = &self.sort {
+            self.apply_sort(&selected, sort)?
+        } else {
+            selected
+        };
+
+        // Apply pagination
+        let paginated = self.apply_pagination(&sorted);
+
+        // Return count if requested
+        if self.count_only {
+            return Ok(paginated.len().to_string());
+        }
+
+        // Format output
+        let formatter = DefaultFormatter;
+        formatter.format(&paginated, self.format)
+    }
+
+    fn apply_filter(&self, data: &[JsonValue], filter: &Filter) -> Result<Vec<JsonValue>> {
+        let mut result = Vec::new();
+
+        for item in data {
+            if filter.evaluate(item)? {
+                result.push(item.clone());
+            }
+        }
+
+        Ok(result)
+    }
+
+    fn select_fields(&self, data: &[JsonValue], selector: &FieldSelector) -> Result<Vec<JsonValue>> {
+        let mut result = Vec::new();
+
+        for item in data {
+            if let JsonValue::Object(map) = item {
+                let mut new_map = serde_json::Map::new();
+
+                if let Some(include) = &selector.include {
+                    for field in include {
+                        if let Some(value) = map.get(field) {
+                            new_map.insert(field.clone(), value.clone());
+                        }
+                    }
+                } else if let Some(exclude) = &selector.exclude {
+                    for (key, value) in map {
+                        if !exclude.contains(key) {
+                            new_map.insert(key.clone(), value.clone());
+                        }
+                    }
+                } else {
+                    new_map = map.clone();
+                }
+
+                result.push(JsonValue::Object(new_map));
+            } else {
+                result.push(item.clone());
+            }
+        }
+
+        Ok(result)
+    }
+
+    fn apply_sort(&self, data: &[JsonValue], sort_keys: &[SortKey]) -> Result<Vec<JsonValue>> {
+        let mut result = data.to_vec();
+
+        result.sort_by(|a, b| {
+            for sort_key in sort_keys {
+                let a_val = self.get_field_for_sort(a, &sort_key.field);
+                let b_val = self.get_field_for_sort(b, &sort_key.field);
+
+                let cmp = match (&a_val, &b_val) {
+                    (Some(JsonValue::Number(a)), Some(JsonValue::Number(b))) => {
+                        a.as_f64().partial_cmp(&b.as_f64()).unwrap_or(std::cmp::Ordering::Equal)
+                    }
+                    (Some(JsonValue::String(a)), Some(JsonValue::String(b))) => a.cmp(b),
+                    _ => std::cmp::Ordering::Equal,
+                };
+
+                let final_cmp = if sort_key.descending {
+                    cmp.reverse()
+                } else {
+                    cmp
+                };
+
+                if final_cmp != std::cmp::Ordering::Equal {
+                    return final_cmp;
+                }
+            }
+
+            std::cmp::Ordering::Equal
+        });
+
+        Ok(result)
+    }
+
+    fn get_field_for_sort(&self, value: &JsonValue, field: &str) -> Option<JsonValue> {
+        if let JsonValue::Object(map) = value {
+            map.get(field).cloned()
+        } else {
+            None
+        }
+    }
+
+    fn apply_pagination(&self, data: &[JsonValue]) -> Vec<JsonValue> {
+        let offset = self.offset.unwrap_or(0);
+        let limit = self.limit.unwrap_or(usize::MAX);
+
+        data.iter()
+            .skip(offset)
+            .take(limit)
+            .cloned()
+            .collect()
+    }
+}
+
+pub struct FieldSelector {
+    pub include: Option<Vec<String>>,
+    pub exclude: Option<Vec<String>>,
+}
+
+impl FieldSelector {
+    pub fn include(fields: Vec<String>) -> Self {
+        Self {
+            include: Some(fields),
+            exclude: None,
+        }
+    }
+
+    pub fn exclude(fields: Vec<String>) -> Self {
+        Self {
+            include: None,
+            exclude: Some(fields),
+        }
+    }
+
+    pub fn parse(input: &str) -> Result<Self> {
+        if input.starts_with('-') {
+            // Exclude fields
+            let fields: Vec<String> = input
+                .trim_start_matches('-')
+                .split(',')
+                .map(|s| s.trim().to_string())
+                .collect();
+            Ok(Self::exclude(fields))
+        } else {
+            // Include fields
+            let fields: Vec<String> = input
+                .split(',')
+                .map(|s| s.trim().to_string())
+                .collect();
+            Ok(Self::include(fields))
+        }
+    }
+}
+
+pub struct SortKey {
+    pub field: String,
+    pub descending: bool,
+}
+
+impl SortKey {
+    pub fn parse(input: &str) -> Vec<Self> {
+        input
+            .split(',')
+            .map(|s| {
+                let s = s.trim();
+                if s.starts_with('-') {
+                    SortKey {
+                        field: s.trim_start_matches('-').to_string(),
+                        descending: true,
+                    }
+                } else {
+                    SortKey {
+                        field: s.to_string(),
+                        descending: false,
+                    }
+                }
+            })
+            .collect()
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_data_type_parsing() {
+        assert_eq!(DataType::from_str("functions").unwrap(), DataType::Functions);
+        assert_eq!(DataType::from_str("fn").unwrap(), DataType::Functions);
+        assert_eq!(DataType::from_str("strings").unwrap(), DataType::Strings);
+    }
+
+    #[test]
+    fn test_field_selector_parse() {
+        let selector = FieldSelector::parse("name,address,size").unwrap();
+        assert!(selector.include.is_some());
+        assert_eq!(selector.include.unwrap().len(), 3);
+
+        let selector = FieldSelector::parse("-metadata,internal").unwrap();
+        assert!(selector.exclude.is_some());
+    }
+
+    #[test]
+    fn test_sort_key_parse() {
+        let keys = SortKey::parse("name,-size");
+        assert_eq!(keys.len(), 2);
+        assert_eq!(keys[0].field, "name");
+        assert!(!keys[0].descending);
+        assert_eq!(keys[1].field, "size");
+        assert!(keys[1].descending);
+    }
+}