show-observe-run-reportlisted

# 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 report to render, in one of two forms: - **The persistence return value** — what `create-observe-run-report` just returned for the mission being closed, carried in the agent's reply: the stored path, the carrying commit (or `not committed`), and the report as written. Render from it; never re-read the file it just wrote — the value is the file's content, verbatim (the way `show-benchmark` renders from `create-update-benchmark`'s return). - **A stored report the caller names** — no return value in hand: read that file from disk, and its carrying commit from git (`git log -1 --format=%h -- <path>`). Either way the synthesis renders the stored content, never the conversation's memory of the mission. ## 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