ox-recap

---

name: ox-recap description: >- Answer "what value am I getting from SageOx?" with receipts, not vibes. Auto-fire when the user asks "is SageOx worth it", "what has SageOx done for me", "what am I getting out of SageOx", "should I keep using this", "was SageOx worth it this week", or says "ox recap" / "give me a recap". Run ox recap --json and narrate a tight, personal, prose answer from the guidance field — never invent value, never cite a bare statistic or a time-saved/dollar figure, ground every claim in a receipt from the bundle.

<!-- Thin by design. The entire narration contract — which value axis to lead with (social: team knowledge that reached you; temporal/solo: your own ledger compounding as searchable memory), the honesty rules (never invent value, never lead with a bare statistic, ground every claim in a receipt), and the cold-start framing — lives in the `guidance` field of `ox recap --json` (Layer-1 floor), which reaches every primed AI coworker. This native skill adds discovery and activation ergonomics; it duplicates no reasoning. Do not grow this body. -->

Use when

The user asks about SageOx's value, ROI, or whether it's worth keeping: "what value am I getting from SageOx", "is SageOx worth it", "what has SageOx done for me", "what's this actually doing for me", or "ox recap".

Do

1. Run ox recap --json (add --since <window> or --user <name> only if the user asked for a scope other than the default last 30 days).
2. Narrate the answer by following the guidance string in the JSON output verbatim — it decides which value axis leads, enforces the honesty rules, and tells you how to ground every claim in a receipt (an artifact path, a session name, a plan slug, a commit SHA) from the bundle.
3. If the bundle is thin or the ledger is empty, do not invent value — guidance already tells you how to frame that case: lead with next_actions and say plainly that value starts the moment they begin.