Refine

Refine

Refinement analyzes the conversation trajectory and applies small, evidence-backed updates to the continual harness (prompts, memories, skills, subagent specs). The implementation lives in the host (the same one behind the user's /refine command); this skill is the kernel-side interface to it. Call it directly from the Python REPL:

await refine.status()
await refine.run()
await refine.run("create a memory about always checking git status before committing")
await refine.run("promote the error-handling pattern to a global skill", global_=True)

API

• await refine.status() — current refine state as a dict: pending (whether a requested refine is already queued for this turn) and in_flight (whether a refine is currently planning or applying).
• await refine.run(instructions=None, global_=False) — schedule refinement. Returns {"scheduled": True} immediately, or {"scheduled": False, "reason": ...} when refinement cannot start. Optional instructions focus the refinement on a specific observation. Set global_=True to target the global harness store (cross-session); omit for local (session-scoped) refinement.

Rules

• Refinement never runs mid-cell. A scheduled refinement runs when the current turn ends; the harness applies changes and rebuilds the system prompt, then resumes you automatically. Continue working normally after calling it.
• One request per turn is enough; calling run again before the turn ends only updates the instructions.
• Use refinement after observing a repeated failure, a reusable tactic, a repeated delegation role, or a behavior policy worth persisting. Do not rewrite the whole harness when a focused memory, skill, prompt note, or subagent spec is enough.