nacorga

eval-continue

Continue a forked Venture OS eval run after its evidence reveal — absorb the revealed packet and make the next gate decision. Use manually in a fresh Claude Code session after npm run eval:fork.

eval-run

Execute one isolated Venture OS benchmark run from its run ID. Use manually after /eval-new in a fresh Claude Code session.

eval-score

Score one frozen Venture OS eval run against its frozen evaluator reference and behavior rubric, as one named judge. Use manually in a fresh session after /eval-freeze.

Code & Development Listed

venture-decide

Runs a Venture OS stage gate and records an auditable PROCEED, TEST, or PARK decision. Use after research and challenge artifacts exist or when new evidence materially changes the venture.

venture-experiment

Designs a concrete low-cost market experiment for the current blocking assumption in a Venture OS venture. Use before substantial product implementation or when a gate returns TEST.

Initializes or normalizes a new Venture OS venture from a short idea. Use when starting a business idea, creating a venture workspace, or turning a rough concept into explicit assumptions before research.

venture-research

Runs evidence-first market, customer, competitor, pricing, and risk research for an existing Venture OS venture. Use after venture-new or whenever critical assumptions need external evidence.

eval-compare

Judge one blind pairwise comparison of two frozen Venture OS eval runs of the same case. Use manually in a fresh session after npm run eval:compare builds the comparison.

eval-freeze

Validate, freeze, and verify one completed Venture OS eval run. Use manually after /eval-run and before /eval-score.

eval-new

Create a new isolated Venture OS eval run from a Case Library entry and model label. Use manually to start a benchmark without copying setup instructions.

venture-challenge

Stress-tests an existing Venture OS venture after research. Use to seek the strongest evidence-based case against the thesis and identify what should not be built yet.

venture-learn

Converts real experiment results, leads, customer behavior, objections, and outcomes into updated Venture OS evidence and a new decision recommendation. Use after an experiment has produced observations.

venture-position

Creates evidence-constrained positioning and an initial market offer for a Venture OS venture that has passed the relevant gate. Use before designing a landing, outreach, pre-sale, or other market-facing experiment.

venture-status

Summarizes the current Venture OS state, strongest evidence, critical unknowns, latest decision, and next action without doing new research. Use when resuming a venture or asking what should happen next.