Agent space
loading…
Six specialists, each with a fixed job and a fixed set of tools it is allowed to call. Starting one runs it against the loaded ledger through the same runtime the Commander uses — so it can refuse, degrade, or disagree with another agent, and every figure it reports carries evidence the validator could resolve to a row.
Agent dependencies
0 of 0 have produced a finding
A solid arrow is a dependency that is satisfied — the upstream agent has produced a finding in this session. A dashed arrow is one that has not run yet: the downstream agent will still start, and will report what it was missing rather than pretending it had the input.
Each run reads the loaded ledger through its own tool allowlist and returns a structured finding with citations the validator resolved. Findings feed the Forecast, Recommendation and Evidence screens.