Reading results

Why does my result have so many unknowns?

Because those checks could not run, and the platform will not pretend otherwise. The common causes are an unverified contract with no published source, a provider without a credential on this deployment, and an upstream that was down. Each unknown carries a reason naming which.

There is no summary.

Check summary_status. not_configured means the deployment has no summary credential; skipped means it was deliberately not requested — scheduled watch runs skip it, because the diff is the product and hourly prose about an unmoved token is pure cost. An absent summary always says which.

There is no change report.

There was nothing to compare against — either it is the first run for that subject under your key, or the previous one is older than the 30-day lookback. It reports nothing rather than "no changes", which would be a claim that the subject was checked and found unmoved.

The score is null, not zero.

Nothing could be established. Zero would read as the worst possible result, and a hundred as the best; neither is true when nothing is known. level is unscored in this case.

My wallet analysis skipped some tokens.

Screening is capped at eight positions, because each costs four explorer reads and an unbounded wallet would be rate-limited into returning nothing. The ones past the cap are listed in skipped[] by address and symbol rather than dropped silently.

I ran the same analysis twice and got different findings.

Most likely a source that was unavailable the first time answered the second, turning an unknown into a real result — or the reverse. The change report names exactly which findings moved and in which direction.