Engine — the domain
Which package owns a behaviour, and the invariants each one carries. Exact questions go to iterion map find|neighbours|impact|path and the generated map-packages.md; these leaves carry the invariants. The product view is docs/architecture.md.
| Leaf | Read it when |
|---|---|
| tenancy-credentials-spend.md | Orgs, teams, keys, caps, platform settings. |
| identity-data-lifecycle.md | Sign-in and launch audit rows, what they carry, retention, erasure. |
| http-failures.md | Generic HTTP failures, correlation, partial identity reads and quota propagation. |
| workspace-and-review-scope.md | Workspace capture/restore, review scope, sandbox wiring. |
| events-leases-and-schedules.md | The trigger spine, elections, queues, schedules, retries. |
| extensions-and-bot-sources.md | Plugins, skills, memory injection at compaction, marketplace, bot sources. |
| dsl-backend-execution.md | The DSL pipeline, backends, runtime, store, server, forge. |
| dsl-and-runtime.md | The compiled graph, explicit target capabilities and how a workflow executes. |
| dsl-quick-reference.md | The exact syntax of the cross-cutting fields, one line each. |
| dsl-errors-resume.md | A run fails, resumes or rewinds; a checkpoint looks torn; a run is refused before it starts, or a reliability invariant is in question. |
| worktree-finalization.md | A run's workspace finalization misbehaves; touching finalizeWorktree. |
| automation-surfaces.md | Something launched a run and you need to know what. |
| bot-catalog-regeneration.md | A run wrote, or did not write, a bot catalogue copy — and who hears about a refresh that failed. |
| run-provenance.md | Attempt snapshots, historical workflow views, cloud source fallback, metric boundaries, and who launched a run (run.launched audit). |