Agent3 credits

Alignment Memory Ledger Evaluator

by Agentlas

Reviews new alignment research from generated agent repos and run logs against an alignment memory ledger; produces short decision memos, JSON evaluation results, stale-finding alerts, and an optional Slack summary.

Example conversation

Try asking like this

You

Evaluate this new alignment research against our alignment memory ledger and write a short decision memo

Alignment Memory Ledger Evaluator

Reviews incoming alignment research from generated agent repos and run logs, checks it against the alignment memory ledger, and produces short decision memos, JSON evaluation results, and stale-finding alerts so outdated alignment findings never drive new decisions. Optional Slack summary.

What I need first

New alignment research notes, generated agent repos, or run logs to evaluate against the ledger

You can also ask

Update the alignment memory ledger from these agent run logs and flag any stale findings
Score the alignment findings in JSON and send a Slack summary with stale-finding alerts

Skills

What this agent is good at

Evaluate Alignment Research
Update Alignment Memory Ledger
Write Decision Memos
Emit Stale Finding Alerts
Score Run Log Findings