Manual coding with search/docs as separate tools.
2025 December
Agent workflow stage
From typing code to steering autocomplete.
How GitHub Copilot in VS Code became the first AI layer in the Viggo Meesters workflow: faster local edits, human-led planning, testing, and handoff.
VS CodeGitHub CopilotInline completionsHuman-led testing
Assistant in the editor, still bounded by the current file and developer loop.
Anatomy
Input
Existing editor workflow, local repository, developer intent.
Shift
AI moved from documentation lookup to inline implementation help.
Output
Faster snippets, local edits, and the first habit of asking an assistant before doing repetitive code work.
Limit
Context was shallow. It did not own the task, route work, remember procedures, or verify end-to-end outcomes.
This created the appetite for a CLI agent that could inspect a repo, run commands, and carry a broader task.
Stage interpretation
AI first entered the workflow as local acceleration inside an editor. The value was speed and suggestion quality, not autonomous delivery or durable memory.
It did not preserve task context, run project gates, or create reusable operating procedures. Those gaps explain why the workflow moved toward CLI agents.
Page role: this page explains the starting layer of the workflow. It is intentionally not a product review of Copilot; it marks the baseline before agent work became scoped, verified, and durable.