Check the run before the agent starts.
AssetMason turns a task into a reviewable plan, a bounded lock, and an evidence-backed receipt. Your project owns the work. The agent is replaceable.
Use it with your current coding agent. No hosted dashboard required for the individual workflow.
Add audit logging to checkout
A review sequence, not a dashboard
Make the boundary legible before work begins.
Task
Start with the work you actually need done.
Context
Ground the agent in the project and source-linked facts.
Run Plan
Review reuse, additions, exclusions, risks, and proof.
Lock → Receipt
See the allowed boundary and the evidence that came back.
Source-linked result
No addition required is a successful result.
A good plan can say the project already has the right boundary. AssetMason makes that reasoning visible instead of making work look valuable just because it is new.
Reused existing receipt projection.
Evidence: src/lib/assetmason/services/receipt-handoff.ts
Run your first local scan
Keep the project in control.
Install the published preview CLI, scan the repository, and review source-linked output before an agent acts. This hosted fixture is explanatory; it is not live verification.
npx -y assetmason-cli@preview scan --root . --format markdown