--- title: "Vitaly D. (@t3chn) | Hashnode" description: "Vitaly D. (@t3chn) writes on Hashnode. Read their latest articles and follow their work." url: "https://hashnode.com/@t3chn" --- @t3chn # Vitaly D. @t3chn · Joined March 2026 ### [Signum Can Now Be Installed in Codex App as a Plugin](https://t3chn.hashnode.dev/signum-can-now-be-installed-in-codex-app-as-a-plugin?utm_source=hashnode&utm_medium=feed) AI-agent workflows have a boring but important problem: a good process often lives inside one runtime. Signum started as a Claude Code plugin. That was the right first layer: slash command, local artifacts, contract-first pipeline, proofpack. But the... 4 min read · [t3chn.hashnode.dev](https://t3chn.hashnode.dev/signum-can-now-be-installed-in-codex-app-as-a-plugin?utm_source=hashnode&utm_medium=feed) [May 8](https://hashnode.com/posts/signum-can-now-be-installed-in-codex-app-as-a-plugin/69fe085a4a13089bbb93b6a9) ### [Stop Writing CLAUDE.md From Scratch](https://t3chn.hashnode.dev/stop-writing-claudemd-from-scratch?utm_source=hashnode&utm_medium=feed) Every agentic codebase starts with an empty CLAUDE.md and fills it with hallucinations. Here's the six-file harness signum scaffolds for you. 6 min read · [t3chn.hashnode.dev](https://t3chn.hashnode.dev/stop-writing-claudemd-from-scratch?utm_source=hashnode&utm_medium=feed) [Apr 27](https://hashnode.com/posts/stop-writing-claudemd-from-scratch/69ef813d518b04856ce4937a) ### [AI Agents Need Permission Boundaries, Not Personalities](https://t3chn.hashnode.dev/ai-agents-need-permission-boundaries-not-personalities?utm_source=hashnode&utm_medium=feed) Most agent tooling mistakes coordination for reliability. It gives you more roles, more agents, more orchestration, and more shell theater. The demo gets more impressive. The system does not necessarily get easier to trust. That tradeoff used to be t... 7 min read · [t3chn.hashnode.dev](https://t3chn.hashnode.dev/ai-agents-need-permission-boundaries-not-personalities?utm_source=hashnode&utm_medium=feed) [Apr 8](https://hashnode.com/posts/ai-agents-need-permission-boundaries-not-personalities/69d685a8c9c19091ef75ba55) ### [My AI Agent Said 'Done.' It Skipped an Entire Acceptance Criterion.](https://t3chn.hashnode.dev/my-ai-agent-said-done-it-skipped-an-entire-acceptance-criterion?utm_source=hashnode&utm_medium=feed) The hardest bug was not in the code. It was in the trust model between the engineer agent and the orchestrator. 5 min read · [t3chn.hashnode.dev](https://t3chn.hashnode.dev/my-ai-agent-said-done-it-skipped-an-entire-acceptance-criterion?utm_source=hashnode&utm_medium=feed) [Mar 23](https://hashnode.com/posts/my-ai-agent-said-done-it-skipped-an-entire-acceptance-criterion/69c0fab761ba27cd4a6deb11) ### [Your AI Spec Is Already Stale](https://t3chn.hashnode.dev/your-ai-spec-is-already-stale?utm_source=hashnode&utm_medium=feed) When agents read project.intent.md as ground truth, stale specs become execution bugs. Here's how I caught real drift in two projects. 5 min read · [t3chn.hashnode.dev](https://t3chn.hashnode.dev/your-ai-spec-is-already-stale?utm_source=hashnode&utm_medium=feed) [Mar 20](https://hashnode.com/posts/your-ai-spec-is-already-stale/69bd07cfe40e5b04d5ef6aa4)