Proposal: add an agent layer for artifact creation

[clay-good]

Note

Moved from issue #593, originally opened by @roeehersh on 2026-01-26. Continuing the conversation here in Discussions — the original issue thread (including any comments) stays available at #593 for the record.

---

when you have a long exploratory conversation before formalizing a change, the full history gets passed to the command including abandoned ideas and back-and-forth. This bloats the context and can confuse the output.

After exploring with the user, Claude summarizes the relevant context and spawns an artifact-writer agent.
The agent uses existing commands (/opsx:continue, etc.) to create the artifact - but with clean, distilled context instead of the full
conversation history.

Commands stay as-is. The agent is just a context-filtering layer that delegates to them.