Claude Code CLI - Agentic Search

Overview

This example demonstrates the agentic capabilities of Claude Code CLI by having it search through the codebase to find and explain validation examples. The AgentCtrlConsoleLogger provides full visibility into tool calls, streaming, and the agent's decision-making process.

Key concepts: - AgentCtrl::claudeCode(): Factory for Claude Code agent builder - withMaxTurns(): Allow multiple turns for exploration - onText() / onToolUse(): Real-time streaming callbacks - AgentCtrlConsoleLogger: Shows execution lifecycle alongside streaming output

Example

<?php
require 'examples/boot.php';

use Cognesy\AgentCtrl\AgentCtrl;
use Cognesy\AgentCtrl\Broadcasting\AgentCtrlConsoleLogger;

// Console logger for execution lifecycle visibility
$logger = new AgentCtrlConsoleLogger(
    useColors: true,
    showTimestamps: true,
    showToolArgs: true,   // Show tool input args
    showSandbox: true,    // Show sandbox setup
);

$toolCalls = [];

echo "=== Agent Execution Log ===\n\n";

$response = AgentCtrl::claudeCode()
    ->wiretap($logger->wiretap())
    ->withMaxTurns(10)
    ->onText(function (string $text) {
        echo $text;
    })
    ->onToolUse(function (string $tool, array $input, ?string $output) use (&$toolCalls) {
        $target = $input['pattern'] ?? $input['file_path'] ?? $input['command'] ?? '';
        if (strlen($target) > 50) {
            $target = '...' . substr($target, -47);
        }
        $toolCalls[] = $tool;
        echo "\n  >> [{$tool}] {$target}\n";
    })
    ->executeStreaming(<<<'PROMPT'
Complete this task in steps:
1. Use Glob or find command to locate PHP files with "validation" in the filename under ./examples
2. Read the contents of relevant PHP file
3. Analyze the code and provide a concise explanation (under 200 words) covering:
   - What validation is being performed
   - What validation constraints/attributes are used
   - How validation is triggered
   - What happens when validation fails

Provide your final explanation as a clear, structured response.
PROMPT);

echo "\n=== Result ===\n";
echo "Tools used: " . implode(' > ', $toolCalls) . "\n";
echo "Total tool calls: " . count($toolCalls) . "\n";
echo "Exit code: {$response->exitCode}\n";

if ($response->usage) {
    echo "Tokens: {$response->usage->input} in / {$response->usage->output} out\n";
}
if ($response->cost) {
    echo "Cost: $" . number_format($response->cost, 4) . "\n";
}

if (!$response->isSuccess()) {
    echo "Error: Agent search failed with exit code {$response->exitCode}\n";
    exit(1);
}
?>

Expected Output

=== Agent Execution Log ===

14:32:15.123 [claude-code] [EXEC] Execution started [prompt=Complete this task in steps:...]
14:32:15.234 [claude-code] [SBOX] Policy configured [driver=host, timeout=120s, network=on]
14:32:15.235 [claude-code] [SBOX] Ready [driver=host, setup=1ms]
14:32:15.236 [claude-code] [PROC] Process started [commands=12]
I'll search for PHP files with "validation" in the filename under ./examples.
14:32:16.456 [claude-code] [TOOL] Glob {pattern=./examples/**/*validation*.php}

  >> [Glob] ./examples/**/*validation*.php

Found a validation example. Let me read it.
14:32:17.234 [claude-code] [TOOL] Read {file_path=./examples/A01_Basics/Validation/run.php}

  >> [Read] ./examples/A01_Basics/Validation/run.php

The validation example demonstrates...
14:32:19.890 [claude-code] [DONE] Execution completed [exit=0, tools=2]

=== Result ===
Tools used: Glob > Read
Total tool calls: 2
Exit code: 0

Key Points

• Agentic search: Agent autonomously explores files and synthesizes answers
• Full visibility: Console logger shows every tool call alongside streaming output
• Sandbox awareness: Enable showSandbox to see sandbox initialization details
• Multi-turn: withMaxTurns(10) allows the agent to explore iteratively