> ## Documentation Index
> Fetch the complete documentation index at: https://docs.oneshotagent.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Shadow score and ranked review

> Use evidence-based priority without mistaking it for a prediction.

A prospect priority score summarizes available evidence on a 0–100 scale. It is a deterministic heuristic, not a model's probability that someone will buy.

The components are person fit, account fit, intent strength, timing freshness, signal confidence and contactability. The implementation weights and combines them, rounds the result, and records reasons alongside the score. Missing evidence is neutral rather than automatically bad.

## Shadow versus ranked

A badge labelled **shadow** is informational. It lets you inspect the scoring experiment without treating the score as a sending decision.

Queue also supports **ranked** review ordering. When selected, scores help order the review list, interleaved across finders with exploration slots. Newest ordering remains available. This ordering choice does not itself approve a person or send a message.

## Read the reasons

Expand a row to inspect its evidence and components. A high score cannot rescue an off-ICP, duplicate or undeliverable candidate. A low score is not a rejection: check whether the record lacks evidence, or whether the evidence actually argues against fit.

Scores carry a version because heuristics can change. Use `find score-prospects` to backfill scoring and `find calibrate` to inspect calibration options; consult the [generated CLI reference](/oneshot-gtm/cli-reference) before running them. Neither a polished score nor a senior title substitutes for reviewing what the person is building.