# Weekly Agent Review Template

Use this once a week to convert agent activity into compounding memory, constraints, and workflow improvements.

```md
# Weekly Agent Review

Week of:
Reviewer:

## 1. Outputs

List what agents produced this week.

- Drafts:
- Research notes:
- Code or operational changes:
- Published posts:
- Plans or decisions:
- Other artifacts:

## 2. What actually mattered?

Separate motion from value.

- What advanced a real goal?
- What clarified a decision?
- What improved a workflow?
- What created reusable context?

## 3. What context was missing?

Identify context debt.

- Missing project brief:
- Missing source list:
- Missing standards:
- Missing decision history:
- Missing examples:
- Missing constraints:

## 4. What should become a file?

Turn repeated context into durable infrastructure.

- Decision log update:
- Source map update:
- Workflow update:
- Checklist update:
- Standard update:
- Agent role update:

## 5. What should be constrained?

Add judgment gates where autonomy could drift.

- Actions requiring approval:
- Claims requiring verification:
- Areas to avoid:
- Safety or quality gates:

## 6. What is the next experiment?

Choose one improvement for next week.

- Experiment:
- Why it matters:
- How to evaluate it:
- File/workflow to update after completion:
```

## Rule

Do not end the review with observations only. End by updating at least one durable artifact: a memory file, workflow, checklist, source map, decision log, agent role, or constraint.
