November 2025 Summaries

In a recent security incident, PostHog's JavaScript SDKs were compromised by the Shai-Hulud 2.0 worm, which exploited a vulnerability in their workflow automation to publish malicious packages on npm. The attack involved stealing a GitHub Personal Access Token from one of PostHog's bots, which allowed the attacker to exfiltrate sensitive credentials and propagate the breach. PostHog quickly responded by identifying and deleting the malicious packages, revoking tokens, and rolling potentially compromised credentials. The attack highlighted a misunderstanding in PostHog's use of GitHub's workflow triggers, allowing a pull request to execute arbitrary code and steal credentials. In response, PostHog has enhanced its security measures, including tightening package release workflows, increasing scrutiny on workflow modifications, and improving GitHub secrets management. This incident has prompted PostHog to prioritize broad security measures and seek talent to strengthen their security team.

PostHog AI, an advanced AI agent developed by PostHog, has transformed from a simple trend chart tool into a comprehensive analytics assistant capable of accessing and processing data across various tools within the platform. It performs multi-step analyses, writes SQL queries, sets up feature flags, and delves into error diagnostics, improving productivity and efficiency for users. The development journey highlighted challenges such as model improvements, which frequently altered capabilities, and the realization that single-loop architectures were more effective than graph-based workflows or specialized subagents due to context retention. Key insights included the importance of maintaining context, the utility of to-do lists for task tracking, and the benefits of transparent processing for user trust. PostHog AI's architecture emphasizes flexibility and adaptability, avoiding the pitfalls of restrictive frameworks, and focuses on real-world usage evaluations to enhance its functionality. Looking ahead, PostHog plans further advancements in research capabilities, session analysis, and proactive insights, continuing to refine the AI's integration with code and expand its analytical prowess.

PostHog's AI journey began with a chatbot named Max, which evolved into a product assistant, but struggled to break free from being perceived merely as a chatbot. This perception issue led PostHog to retire Max in favor of integrating AI more deeply as an architectural element of their product intelligence, transforming it from a novelty into a core system that enhances user experience. The overhaul of PostHog AI's infrastructure improved its capabilities, enabling more reliable and faster responses, reduced errors, and better contextual understanding. This shift allowed PostHog AI to support technical users by automating routine tasks, providing insights, and improving decision-making processes, which in turn facilitated broader adoption across organizations. The move from an assistant to a system reflects PostHog's commitment to building an AI that is integral to its platform, enhancing functionality and user workflows without relying on a mascot-driven experience.

PostHog's Workflows, now in beta after significant improvements from its alpha phase, allows users to automate product-led actions like sending emails, updating user properties, or triggering Slack alerts directly from the tracked PostHog event data without the need for third-party integrations. This feature, reminiscent of tools like Zapier and Active Campaign, facilitates faster and more reliable automation by eliminating the need for external data syncing, offering capabilities such as branching logic and live event-triggered actions. Users, including teams like Grantable, have found it significantly more efficient, with the setup being easy and almost twice as fast compared to other tools, as noted by Evan Rallis, who appreciates the streamlined process of building automations directly within PostHog. The beta version remains free, encouraging users to experiment with small-scale workflows to enhance their product's interactivity and responsiveness.

Rollbar is a popular error tracking tool known for its simplicity and real-time error detection capabilities, catering primarily to developers needing reliable error alerts. However, as products grow, there's a need for more comprehensive solutions that offer context beyond basic error reporting, such as session replay, debugging context, and observability. Alternatives like PostHog, Sentry, LogRocket, Datadog, Bugsnag, and Airbrake provide expanded capabilities. PostHog, for example, integrates error tracking with product analytics and session replays, while Sentry offers performance monitoring and deep stack traces. LogRocket focuses on frontend debugging with session replays, Datadog provides full-stack observability, Bugsnag excels in mobile app stability, and Airbrake offers a straightforward and affordable solution for smaller teams. These alternatives address various needs, from comprehensive user session insights to enterprise-scale observability, with options for different budgets and team sizes.

PostHog's Workflows is an automation tool that allows users to create and run actions based on live user behavior, offering features like triggers, logic, and actions without the need for syncing or external connectors. Built directly into PostHog, it leverages real-time product data to enable seamless automation, with current support for email and upcoming integrations with platforms like Slack and Twilio. The system's simplicity allows even non-technical users to build automations visually and publish them easily. A product marketer's test of the tool revealed the importance of setting entry conditions to prevent unintended repetitive actions, as demonstrated when a misconfigured workflow led to multiple email triggers for users. Despite this hiccup, the feedback received was overwhelmingly positive, highlighting the ease of use and effectiveness of the tool, ultimately turning the incident into a learning opportunity and a chance to engage with enthusiastic users. Currently in its alpha stage and moving to beta, the tool is available for free trials, inviting users to experiment while being mindful of entry settings.