The protocol for ephemeral decisions

Software engineering has a documentation problem that is only getting worse as development cycles accelerate. While code is version-controlled in Git and tasks are tracked in Jira or Linear, the actual rationale behind a specific choice—the "why" of a product decision—is often lost in the noise of a Slack thread or a local terminal session. Lock is a decision protocol designed to bridge this gap by capturing these moments at the source. It operates on the premise that decisions are the most important units of work in a product organization, yet they are the ones most likely to be poorly documented.

Lock functions as a lightweight layer that sits where engineers and product managers already work. By providing simple commands to log a decision from the CLI or within a Slack conversation, it creates a permanent, searchable record of the path taken. This approach addresses the friction of traditional documentation, which usually requires moving to a separate tool like Notion or a Wiki, a step that is often skipped in the heat of a deployment or a debugging session.

Capturing the intent of AI agents

The introduction of AI agents into the development workflow adds a new layer of complexity. When a human developer makes a change, there is at least a theoretical trail of communication. When an AI agent executes a multi-step workflow—modifying files, updating configurations, or deploying code—it often does so without leaving a clear narrative of its reasoning. Lock is specifically designed to integrate into these agent sessions. By forcing or enabling agents to log their decisions through the Lock protocol, organizations can maintain a human-auditable trail of what an agent did and why it chose that specific action.

This makes Lock a part of the emerging "agentic governance" stack. As companies move from using AI as a chatbot to using it as an autonomous operator, the need for traceability becomes a requirement rather than a feature. If an agent decides to refactor a database schema or change a pricing tier in a config file, Lock provides the mechanism to capture that decision as an event, making it as visible and traceable as a manual commit from a senior engineer.

Market position and utility

Lock occupies a niche between project management and system observability. It is not a task tracker, nor is it a monitoring tool for system performance. Instead, it is an audit log for human and machine intent. In a market where distributed teams and autonomous agents are becoming the norm, the value of a centralized decision store increases. The protocol is built to be minimal, avoiding the bloat of traditional enterprise software. It focuses on the single act of capturing a decision with "one command."

The company targets product teams who feel the pain of lost context. This is particularly relevant in high-growth startups or organizations with high engineer turnover, where "archaeology" is often required to understand why a legacy system was built in a certain way. By integrating directly into the terminal and Slack, Lock attempts to become an invisible part of the developer's muscle memory, ensuring that the history of a product is written in real-time rather than reconstructed from memory weeks later.