‹ IT & Security recipesIT & Security · Agent Blueprint
Change rationale
Why a production change was made. The intent behind every config drift, so the next engineer does not revert a decision the org already paid to learn.
Why this is memory-only. Configuration history shows what changed, not why. The intent and tradeoff usually live in the conversations around the change and disappear when the engineer leaves. A persistent memory layer preserves that rationale.
Side by side: token usage, with and without Paperbase
Memory onlyWithout Paperbase
Not recoverable from configuration history alone
- Inspect the current configuration and git diff.
- Search tickets and change requests for the reason.
- Ask the engineer who made the change.
- Guess whether drift is intentional or accidental.
With Paperbase
Persistent memory layer
- Pull the Decision and Rationale behind the change.
- Pull the interactions that led to it, in chronological order.
- Match the change to its owner, system, and affected commitments.
- Flag changes with no rationale or stale rationale.
Agent prompt
You are investigating the rationale for production change {{change_or_system}}.
Using Paperbase:
1. Find the Decision and Rationale behind {{change_or_system}}.
2. Pull the interactions that led to the change, in chronological order.
3. Identify who approved it, what risk it addressed, and what it affected.
4. Flag whether the rationale is still current.
Output:
- Why the change was made, with provenance.
- The alternatives and risks discussed.
- Whether to preserve, revise, or reverse it, with the evidence.Placeholders in {double_braces} are inputs the agent will ask for at runtime. Give the prompt to any agent connected to Paperbase, and the rest grounds in your own memory graph.