AI Work Library
Good AI work
leaves a trail.
The workflows, instructions and checks behind the builds. Open an example, make it yours, and take it with you.
Open the working notes
Find your next building block.
-
Workflows · Workflow specification
Investigate before you escalate
Bring renewal, usage and support evidence into one reviewable customer-risk decision.
-
Workflows · Workflow specification
Read the customer, then the queue
Triage a case against account context and what the customer actually said.
-
Agents · Agent instruction
A QA agent with a stopping point
Give a reviewer a narrow job, an evidence format and an explicit permission boundary.
-
Agents · Agent instruction
Keep the disagreement visible
A critic role that challenges a recommendation without erasing the original evidence.
-
Prompts · Prompt
Give the idea a buildable brief
Turn rough notes into a bounded first build with explicit unknowns and acceptance criteria.
-
Prompts · Prompt
Leave the next builder a trail
Convert a build session into a handoff that separates shipped work, evidence and remaining risk.
-
Skills · Portable procedure
Make the product explain itself
A reusable procedure for turning a working flow into a clear, inspectable walkthrough.
-
Evaluations · Runnable fixture + specification
Did the evidence actually show up?
Run a small citation-coverage check, then break the fixture to see what fails.
-
Evaluations · Runnable fixture + specification
A proposal is not permission
Test an explicit approval gate before a proposed operation can become eligible.
No matching examples.
Try a broader phrase or another asset type.
Put the pieces together
See the whole loop run.
Investigate a synthetic customer, inspect the evidence, and review a proposed action in the Enterprise Lab.