show-observe-run-report

name	show-observe-run-report
description	Render a short synthesis of a persisted observation report for the human closing the mission - verdict-first headline, the stored path, a compact run block, the findings that matter, and the recommended next action - never a replacement for the report itself. Use when an /odd-observe or /odd-verify mission ends and the final answer must synthesize the stored report instead of dumping it raw.

Show an Observe-Run Report

The stored report is the ODD loop's memory and the fix plan's input — the right artifact for the next wave, the wrong one for the human closing the mission: several screens deep, the takeaways drown. This skill renders the closing synthesis. The report file stays the deliverable; only what the human sees at the end of the mission changes.

Input

The persisted report to render: the file the create-observe-run-report skill just stored, or any stored report the caller names. Read it from disk — the synthesis renders the stored file, not the conversation's memory of it.

The synthesis, in order

1. Headline — one bold line answering "how did it go", shaped by the report's mode: an observation leads with counts and the baseline delta (3 anomalies (1 high, confirmed), 2 telemetry gaps, p95 stable vs baseline); a verification leads with the ruling (FAIL — 2/5 checks red); a re-measure leads with drift (no drift — 5/5 measurements within range).
2. Where it lives — one line: the stored path and the commit that carries it.
3. Run block — compact key: value lines: services, stack, mode, window, detected environment (all from the frontmatter), and the baseline report used — verifies when present, else section 1's recalled baseline, or "none" when the report names none.
4. The core, by kind — tables, capped at ~10 rows with a +N more in the report marker:
   • observation / re-measure: the findings table (severity | confidence | one-line anomaly), then the telemetry gaps, one line each;
   • verification: the verdict table first (check | before | after | pass/fail), then the anomalies ruled fixed / still present / worse and the gaps ruled filled / still missing, one line each — for an instrumentation baseline, the presence rulings instead (planned item | closed / still missing), and nothing else to rule.
5. Decisions the spec must settle — the count, then one line per open question.
6. Next action — one line naming the loop's next step: build the fix plan from the report, replay the protocol with /odd-verify, or settle the open decisions first.

Rules

• Everything comes from the stored report: no backend query, no re-derivation, no invented number (the carrying commit, read from git, is the one value outside the file) — a value the report does not carry is absent from the synthesis too.
• One screen, hard cap: prefer dropping rows (behind the +N more marker) over growing sections.
• Render in the conversation's language; the stored report itself stays English.
• The synthesis never replaces the report: the next wave consumes the stored file — state its path, never re-inline the full body.