The latest articles on DEV Community by moriwo-dev-ai (@moriwodevai). https://dev.to/moriwodevai https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4043550%2Fe08b8e19-09bf-48d3-87ee-1f25d5c86e19.png https://dev.to/moriwodevai en moriwo-dev-ai Thu, 23 Jul 2026 10:02:47 +0000 https://dev.to/moriwodevai/i-built-a-desktop-ai-agent-that-writes-its-own-tools-behind-verification-gates-2jfl https://dev.to/moriwodevai/i-built-a-desktop-ai-agent-that-writes-its-own-tools-behind-verification-gates-2jfl <p><em>A desktop AI agent that writes its own tools — and why every line passes a gate before it runs.</em></p> <h2> The itch </h2> <p>Agent capability is quietly becoming synonymous with a specific vendor's pricing plan. I wanted the opposite: an agent that runs on my machine, where models are swappable parts (Anthropic / OpenAI / Moonshot, or free-tier OpenAI-compatible APIs), and where the tools it writes stay on my PC as my assets.</p> <p>So I built <strong>AMA-teras</strong> (AGPL, Electron + TypeScript): a desktop agent that, when it hits a capability it lacks, writes a new tool plugin for itself.</p> <h2> The scary part, handled first </h2> <p>"Self-evolving" sounds like a horror story, so the pipeline is the product:</p> <ol> <li>Generation happens in an <strong>isolated git worktree</strong> — the running app is never touched</li> <li>The generated tool must pass <strong>typecheck → real unit tests → a smoke run</strong> </li> <li>Then a human reads the full diff and <strong>approves or rejects</strong> </li> <li>Promotion is a git tag; failed health checks <strong>auto-rollback</strong> </li> <li>The agent cannot modify its own approval mechanism or evolution engine (protected paths are enforced by a tripwire gate)</li> </ol> <p>Nothing ships unreviewed — including code it wrote for itself, and anything imported from the community registry (shared plugins carry a verification-evidence record; audits elsewhere report 12–20% of marketplace skills are malicious, so evidence-or-nothing felt right).</p> <h2> What it can actually do </h2> <ul> <li>Ask for a feature in chat ("I want a YAML→JSON tool") → it plans, writes the plugin + tests, runs the gates, asks for approval, and the tool is usable immediately</li> <li>Approve everything from your phone (worktree promotion, publishing, even Bluesky posts go through the same approval gate)</li> <li>Nightly autonomous mode stacks its work on a separate branch for morning review — main is never touched unattended</li> </ul> <h2> Honest limitations </h2> <ul> <li>Windows installer only today (source build elsewhere is untested)</li> <li>The installer is unsigned (SmartScreen warning)</li> <li>Self-evolution is deliberately scoped to tool plugins, not the core</li> </ul> <p>A large part of this app was built by the agent itself, under the same gates. The failure stories are documented in the repo — including the night it deleted its own work branch and we had to rebuild two tools from scratch (that incident is why nightly work now auto-backs-up to a private remote).</p> <p>Repo: <a href="https://github.com/moriwo-dev-ai/ama-teras" rel="noopener noreferrer">https://github.com/moriwo-dev-ai/ama-teras</a></p> <p>Feedback very welcome, especially on the safety model.</p> ai opensource electron typescript