Skip to main content

Replacement Holds

A replacement hold is a safety brake. When a replacement looks risky — for example a previous attempt failed in a way that could waste spend or strand a route — Liskov derives a hold that blocks further resume or replacement until you clear it deliberately.

A hold is not an error. It is Liskov refusing to spend again until a human confirms it is safe.

Recognising A Hold​

proof liskov application status my-app reports the hold and the reason it was derived (for example, a prior execution that did not complete cleanly). No replacement will launch while the hold stands.

Before You Override​

Confirm the previous attempt is genuinely resolved:

  • Read application plans my-app and the latest execution's diagnosis.
  • Confirm no replacement is mid-flight and no route is half-open.
  • Confirm the budget is intact (a partial spend may have already happened).

If you are unsure, treat the hold as correct and investigate with Recovery first.

Clearing It​

Override the hold with an explicit reason, and re-confirm spend:

proof liskov custody execution run-one my-app \
  --override-replacement-hold --reason "prior attempt cancelled before funding" \
  --yes-spend

The --reason is recorded with the execution. Overriding without --yes-spend still will not spend — both gates are required to launch.