Step 1 of 4
Write the memorability criteria
The gate is a checklist, not a vibe.
Define the write gate. MEMORABLE: durable project facts, user preferences stated or clearly implied, decisions with reasons, working recipes verified by success, failures with diagnosed causes. NOT MEMORABLE: ephemeral state (current file open), one-off values, anything reconstructible from the code in seconds, speculation, and secrets (hard block: credentials never enter memory). For each memorable type define required fields (e.g. a recipe requires the command AND the context it worked in). Encode this as the memory tool's validation spec and as guidance in the system prompt. Test the gate with 10 candidate writes (5 should pass, 5 should be rejected) and verify each outcome.
Expected after this step
A criteria spec enforced by the tool, validated on 10 candidates.
Should not happen
- ✕Remembering everything, so recall drowns in noise
- ✕No provenance, so bad memories cannot be traced or trusted
- ✕Contradictory facts accumulating until the agent's worldview is incoherent
- ✕Single-observation 'facts' treated with the same trust as confirmed ones
Verify before continuing
Do not move on until every check is true. The complete button stays locked until then.
Do not continue if…
- !Remembering everything, so recall drowns in noise
- !No provenance, so bad memories cannot be traced or trusted
- !Contradictory facts accumulating until the agent's worldview is incoherent
- !Single-observation 'facts' treated with the same trust as confirmed ones
If the AI messes this up
Use this when the AI fakes progress or breaks the feature. It forces a real fix.
The gate passes junk. Make rejection reasons specific in the tool response ('rejected: ephemeral state') so you can audit which criterion is leaking, then tighten it.