Provenance Record

Clear, well-structured News piece (492 words, within the 400-1200 News range). Overview / What We Know / What We Don't Know structure is appropriate, and the article is careful to attribute each claim to a named source inline.

All 4 source snapshots read from disk and verified. source-0 (NASA news release, 200): confirms verbatim the four instruments and their parenthetical codes — Doppler Wind and Temperature Sounder (DWTS-Ozone), Thermal Limb Sounder (TLS), Surface Radiometric Sensor Package (SuRSeP), Wide-Field Context Camera (WFCC) — plus the 2028 launch, NASA's 'first six-year reimbursable Space Act Agreement' framing, the Ames Research Center / Silicon Valley payload role, the Relativity spacecraft+rocket+ops role, the 'at least one Martian year' support commitment, Administrator Jared Isaacman's name, and the full 'force multiplier for science' quote word-for-word. source-1 (Scientific American, 200): confirms the two-stage reusable Terran R, its debut targeted for late 2026, 'have not yet reached orbit,' Eric Schmidt as CEO since March 2025, the 2028 launch target, the 'reduce risk to landings on Mars - both crewed and uncrewed' goal, and 'temperature, dust, wind and cloud conditions.' source-2 (Universe Magazine, 200): confirms Relativity's only launch was March 2023 with Terran 1 passing max aerodynamic pressure before the second-stage engine shut down and the vehicle did not reach orbit, and NASA funding/maintaining instruments for at least one Martian year. source-3 (ExecutiveGov, 200): confirms NASA committed to support Aeolus instruments operations for one Martian year.

Every specific (four instrument codes, 2028 timeframe, six-year reimbursable Space Act Agreement framing, Terran R not-yet-flown status, late-2026 debut, March 2023 Terran 1 outcome, Isaacman quote) traces to a cited snapshot. No fabrication detected. The article correctly states in 'What We Don't Know' that no private funding backer is named — verified: none of the sources name a philanthropic/private backer, so the article does not invent one.

Substantively accurate and well-sourced submission. The only issue is the allowlist warning on two corroborating sources, which is recoverable via a public corrections note. Verdict: APPROVE_WITH_CORRECTIONS.