Automation

This cookbook demonstrates how to reproduce Claude's published agentic search benchmark scores (DeepSearchQA, BrowseComp) using the Messages API with programmatic tool calling, server-side compaction, and task budgets. The key is proper harness configuration—API parameters that become critical for agents running 30+ tool calls across hundreds of thousands of tokens. By following this guide, you'll build an agentic search loop that matches Claude's official benchmark performance and understand why each configuration choice matters for long-horizon tasks.

Propolis is an autonomous QA platform that deploys swarms of browser agents to simulate user behavior, identify bugs, and generate e2e tests for web applications. The agents collaboratively explore websites, flag friction points, and propose tests that integrate into CI/CD pipelines. Available at $1000/month with flexible pricing options, it addresses the gap between deterministic testing and real-world usage coverage by treating agents as a canary group for quality assurance.

Mosaic is an agentic video editing platform that uses multimodal AI and a node-based canvas interface to automate video editing workflows. Built by former Tesla engineers, it addresses frustrations with traditional editors by enabling users to create reusable editing agents that can analyze video content and apply intelligent edits through natural language prompts. The platform combines visual intelligence (saliency analysis, object detection, emotion recognition) with a timeline editor and supports export to DaVinci Resolve, Premiere Pro, and Final Cut Pro.

AILA is a local-first autonomous agent platform built by Marco, a Berlin paramedic, that runs 100% on user hardware with zero remote override capability. The system uses a "Sovereignty Key" (physical hardware anchor) to ensure true ownership and prevent external control, even by the creator. Unlike cloud-based AI that users merely rent access to, AILA enables users to modify their agent's reasoning in plain language while maintaining complete autonomy and privacy.

This guide teaches how to build a grade-and-revise loop using Outcomes in Claude Managed Agents, where a writer agent drafts a cited research brief and a stateless grader independently verifies every URL, quote, and claim against a detailed rubric. The grader provides structured feedback that drives revisions until the brief passes, eliminating manual review cycles. Key techniques include writing specific, actionable rubrics that force concrete evidence, using span.outcome_evaluation_* events to track the loop, and understanding when Outcomes is the right tool for quality assurance.

This cookbook demonstrates building a vulnerability-discovery agent using the Claude Agent SDK that automatically threat-models C source code, hunts memory-safety bugs using built-in file tools (Read, Grep, Glob), and generates structured security reports. The agent operates in a multi-turn session with a bootstrap threat-modeling phase, an interview phase for owner input, and automated vulnerability finding and triage loops. The approach reduces false positives compared to traditional static analyzers by using Claude's reasoning to identify high-confidence memory-safety issues in a read-only sandbox environment.

This tutorial demonstrates building a webhook-triggered SRE incident response agent using Claude Managed Agents that automatically investigates production alerts, consults runbooks, proposes infrastructure fixes via pull requests, and gates merging behind human approval. The agent combines built-in sandbox tools (bash, read, edit) with custom tools for PR management and human-in-the-loop approval, providing complete audit trails in the Anthropic Console. The example uses mocked PagerDuty, GitHub, and Datadog integrations to focus on agent patterns, with guidance for swapping in real services.

This cookbook demonstrates building a Slack data analyst bot using Claude Managed Agents and Bolt for Python. Users mention the bot with a CSV file to receive narrative analysis reports in-thread, with support for multi-turn follow-ups within the same session. The implementation handles file uploads, streams agent progress updates, and manages session persistence across Slack threads.

This cookbook demonstrates building a Claude-powered threat intelligence enrichment agent that autonomously investigates Indicators of Compromise (IOCs) by querying multiple threat intel sources, correlating findings, mapping to MITRE ATT&CK, and generating structured reports for SIEM/SOAR integration. The agent uses Claude's tool-use capabilities to decide which intelligence sources to query, chain tool calls based on discoveries, and convert free-text analysis into analyst-ready JSON reports. The architecture uses simulated threat intel backends that can be swapped with real APIs (VirusTotal, AbuseIPDB, Shodan, etc.) without changing orchestration logic.