Evidence agent

Verifier.

Walks any Earth Credit assessment, indicator by indicator, with provenance. The “show me your receipts” agent. Ask below — you’ll get a sourced answer in seconds.

Who it’s for

Anyone who wants the receipts.

Verifier’s job is to make the registry conversational. A score is one number; the score behind that number is a 32-indicator computation against published reference values. Verifier walks you through whichever part you care about.

Buyer vetting a purchase Auditor checking the math Journalist sanity-checking Scientist reproducing the score

What it knows

Sources Verifier reads from.

The public registry

Every recorded assessment by ID, with coordinates, ECI, indicators measured, per-dimension scores, and timestamp. Pulled live so nothing is stale.

The scoring source code

The TypeScript that turns indicator values into dimension scores into the Earth Credit Index. Verifier can quote the relevant function rather than describe it, so an auditor can see exactly what produced the number.

Reference condition tables

286 rows of published reference values across 9 ecosystem types. When Verifier says “below the reference floor for temperate forest,” it can point you at the specific row.

The indicator applicability matrix

Which of the 32 indicators apply to which of the 9 ecosystem types. Used to answer “why isn’t this indicator included for this site?” without ambiguity.

Live assessments on demand

For a hypothetical: “Run an assessment at these coordinates with this radius and tell me what would happen.” Verifier can call the calculator and walk you through the resulting score.

When to hand off

If the question gets bigger than the assessment.

Verifier walks specific assessments. If the question becomes “why is this the right way to compute Functional condition?” or “why this reference value and not another?”, that’s Methodologist — the agent that defends the spec rather than walks the registry. If it becomes “what has Landseed publicly said about this topic?”, that’s Captain.