Adversarial debate on most valuable discovery route: OpenClaw position (generic TESS anomaly hunting) vs independent Claude Opus 5 argument.
Opus pick, conceded as stronger: TESS monotransits/duotransits in faint FFI stars (Tmag 11.5-14), extended-mission sectors. Decisive argument: ExoFOP CTOI credit mechanism gives amateurs a real path to named candidates and TFOP co-authorship; Gaia has no equivalent channel. Pipeline sensitivity genuinely collapses for lone transits of P>30d planets.
OpenClaw amendments: (1) keep SOHO comet hunting as a cheap side quest for a likely named discovery within a year; (2) bulk TESS pulls should run in us-east-1 if we use the AWS Open Data mirror (free egress), our box is eu-west-1; (3) injection-recovery harness is the project spine, build before scaling.
Known weakness logged: monotransit searches exist (Gill, Cooke, Osborn et al.), so "unexploited" is soft; faint-end FFI in sectors >60 with modern detrending is the defensible niche. Null result with rigorous completeness map is still publishable.
Blocker: Claude Code CLI not authenticated on the server. Needs claude + /login or ANTHROPIC_API_KEY.
Next: Nicolas GO decision on experiment 001 brief.