FN-5901: reap stale mission validator runs
Add self-healing recovery for stale mission validator runs left behind after execution ownership disappears.
- add mission-store support to find and reap stale running validator runs, preserving terminal error status and resetting eligible features to needs_fix
- teach the mission execution loop and self-healing maintenance sweep to skip live validations, reap abandoned runs, record audit events, and avoid double-completing runs
- extend regression coverage, mission docs, architecture notes, and add a published-package changeset for the new recovery behavior
Files changed:
.changeset/fn-5901-validator-run-reaper.md | 7 +
AGENTS.md | 1 +
docs/architecture.md | 2 +
docs/missions.md | 26 ++-
packages/core/src/__tests__/mission-store.test.ts | 99 +++++++++
packages/core/src/mission-store.ts | 91 ++++++++
packages/engine/src/__tests__/mission-execution-loop.test.ts | 232 +++++++++++++++++++++
packages/engine/src/__tests__/reliability-interactions/mission-validator-run-reaper.test.ts | 181 ++++++++++++++++
packages/engine/src/mission-execution-loop.ts | 102 +++++++--
packages/engine/src/runtimes/in-process-runtime.ts | 8 +-
packages/engine/src/self-healing.ts | 22 ++
11 files changed, 746 insertions(+), 25 deletions(-)
Fusion-Task-Id: FN-5901
Fusion-Task-Lineage: 87eb2f3f-fc31-4e0a-b0fc-b771f6dc48a3