Notes on engineering, context engineering, harness engineering, MCP, and AI agents in production.https://victordantas.dev/enhttps://victordantas.dev/en/blog/context-engineering/https://victordantas.dev/en/blog/context-engineering/Most agent failures I have seen in production did not come from the model - they came from the context. Notes on what goes in, what stays out, and why.Thu, 14 May 2026 00:00:00 GMThttps://victordantas.dev/en/blog/mcp-beyond-hello-world/https://victordantas.dev/en/blog/mcp-beyond-hello-world/Almost every MCP tutorial stops at an endpoint that returns the time. The real work starts with auth, pagination, and transparent errors.Thu, 02 Apr 2026 00:00:00 GMThttps://victordantas.dev/en/blog/harness-engineering/https://victordantas.dev/en/blog/harness-engineering/Everyone argues about which model to use. In practice, an agent's reliability is won or lost in the harness: tools, hooks, permissions, and context.Wed, 18 Mar 2026 00:00:00 GMThttps://victordantas.dev/en/blog/claude-code-skills/https://victordantas.dev/en/blog/claude-code-skills/Every automation looks like a potential skill. Almost none is. Heuristics for deciding before you spend an hour writing metadata.Fri, 27 Feb 2026 00:00:00 GMThttps://victordantas.dev/en/blog/hooks-as-contract/https://victordantas.dev/en/blog/hooks-as-contract/Hooks are where the team's politics become code. What needs to run before, after, and around the agent - and what should fail loudly.Mon, 09 Feb 2026 00:00:00 GMThttps://victordantas.dev/en/blog/ai-committee-2026/https://victordantas.dev/en/blog/ai-committee-2026/A raw snapshot of the main discoveries from Bamse's AI Committee: what became obvious, what surprised us, and what is still open.Wed, 21 Jan 2026 00:00:00 GMT