Institution Memory: answers that sharpen per name over time.
The Trust Ledger already records every verdict and the outcome it reconciled against. Institution Memory reads that history for a single name and derives its track record, so a name LyraMind has repeatedly been wrong about is held to a higher bar than the raw data alone would set.
A read over the ledger, per name
For any symbol, Institution Memory returns how many times it has been read, the distribution of past verdicts, the reconciled outcomes (including the name-specific borne-out rate), and whether recent confidence is rising, steady, or falling. It is a pure read: it records nothing and never raises. On a name with no history, it says so plainly rather than inventing a track record.
The moat that compounds
A name whose reconciled outcomes fall below a borne-out threshold, with enough outcomes to be meaningful, is flagged as a weak track record. The Agent Readiness gate consults this flag, so repeated misses on a name raise the bar for future reads on it. Every answer to the ledger makes the memory larger, and every reconciled outcome makes it sharper. It describes the track record of the read itself, never a buy or sell view.
- Per-name history: times seen, first and last seen, verdict distribution
- Reconciled outcomes with a name-specific borne-out rate
- Confidence trend detection: rising, steady, or falling
- Weak-track-record flag when a name's outcomes fall below threshold
- Feeds the Agent Readiness gate to hold weak names to a higher bar
- Pure read over the Trust Ledger: records nothing, never raises
Questions
Where does the history come from?
Entirely from the Trust Ledger, the append-only record of every verdict LyraMind gave and the outcome it was reconciled against. Institution Memory just reads and summarizes it per name.
What makes a track record weak?
A borne-out rate below the threshold with at least a minimum number of reconciled outcomes. That flag raises the bar the readiness gate applies to future reads on the same name.
Does it give a view on the stock?
No. It describes the track record of LyraMind's own reads on the name, not a directional opinion, and never advice.