Contract traceability¶
This index is generated from quality/contracts.toml. The TOML registry is the
machine authority; this page is its review surface.
| Contract | Owner | Risk | Normative source | Executable evidence | Profiles |
|---|---|---|---|---|---|
observability.signals.safe |
observability |
high | docs/guide/observability.md — 1. Safe signal contract | pytest: 2 | pr, main, release |
observability.logging.correlated |
observability |
medium | docs/guide/observability.md — 5. Process-host ownership | pytest: 2 | pr, main, release |
observability.repository.enforced |
observability |
high | docs/guide/observability.md — 6. Family dispositions | pytest: 1; static: 1 | pr, main, release |
missions.environment.pinned |
missions |
high | docs/guide/agent-missions.md — 5. Sandbox and validator protocol | pytest: 3; static: 1 | pr, main, release |
architecture.dependencies.enforced |
app |
high | docs/guide/application-architecture.md — 11. Static enforcement | pytest: 1; static: 1; eval: 2 | pr, main, release |
architecture.v0_6.current_dag |
architecture |
high | docs/guide/application-architecture.md — 13. Current v0.6 architecture | pytest: 1 | pr, main, release |
architecture.migration.family_edges |
architecture |
high | docs/guide/storage-migration.md — 10. Architecture and executable evidence | pytest: 1; static: 1 | pr, main, release |
migration.local_v1.whole_identity |
migration |
high | docs/guide/storage-migration.md — 1. Local v1 profile | pytest: 1 | pr, main, release |
migration.local_v1.fail_closed |
migration |
high | docs/guide/storage-migration.md — 3. Preflight, quiescence, and destination emptiness | pytest: 1 | pr, main, release |
migration.tables.deterministic_evidence |
storage |
high | docs/guide/storage-migration.md — 4. Data-plane transfer | pytest: 2 | pr, main, release |
migration.artifacts.occurrence_relocation |
artifacts |
high | docs/guide/storage-migration.md — 5. Artifact relocation without re-ingestion | pytest: 1 | pr, main, release |
migration.control.exact_activation |
storage |
high | docs/guide/storage-migration.md — 6. Control import and writer fencing | pytest: 1 | pr, main, release |
migration.activities.empty_only |
migration |
high | docs/guide/storage-migration.md — 3. Preflight, quiescence, and destination emptiness | pytest: 2 | pr, main, release |
architecture.protocols.complete |
app |
high | docs/guide/application-architecture.md — 8. Protocol policy and wiring | pytest: 1; static: 1; eval: 1 | pr, main, release |
storage.execution.single_authority |
storage |
high | docs/guide/application-architecture.md — Storage execution authority | pytest: 3; static: 2 | pr, main, release |
runtime.trust.actor_free |
runtime |
high | docs/guide/runtime.md — R3 — Runtime is actor-free | pytest: 1; static: 1; eval: 1 | pr, main, release |
runtime.lifecycle.single_flight_and_drain |
runtime |
high | docs/guide/specification.md — Concurrency Contract | pytest: 3; eval: 2 | pr, main, release |
runtime.lifecycle.retryable_teardown |
runtime |
high | docs/guide/runtime.md — R4 — Async context manager is canonical | pytest: 6 | pr, main, release |
gateway.authorization.rbac |
commands |
high | docs/guide/command-gate.md — 1. The gate model | pytest: 3; static: 1; eval: 5 | pr, main, release |
missions.control.run_rest_surface |
missions |
high | docs/guide/api-layer.md — MissionRun control surface | pytest: 2 | pr, main, release |
commands.identity.idempotent |
commands |
high | docs/guide/specification.md — Command ledger, scheduler, and dispatcher | pytest: 1; static: 1; eval: 4 | pr, main, release |
commands.settlement.atomic |
commands |
high | docs/guide/application-architecture.md — 7. Commands, commits, artifacts, and audit | pytest: 3; eval: 2 | pr, main, release |
commands.failure.preserves_progress |
commands |
high | docs/guide/specification.md — Command ledger, scheduler, and dispatcher | pytest: 1; eval: 4 | pr, main, release |
world.tick.atomic_visibility |
world |
high | docs/guide/application-architecture.md — 10. Supported durability profile | pytest: 1; eval: 2 | pr, main, release |
world.writer.fenced |
world |
high | docs/guide/world-lifecycle.md — 6. resume_world (fenced mutable cold resume) |
pytest: 2; eval: 3 | pr, main, release |
activities.durable_control |
activities |
high | docs/guide/activities.md — 2. The committed-state protocol | pytest: 1; static: 2 | pr, main, release |
missions.activity.recovery |
missions |
high | docs/missions/recovery.md — Mission Activity recovery | pytest: 5 | pr, main, release |
world.lifecycle.idempotent |
world |
high | docs/guide/world-lifecycle.md — 2. World lifecycle operations | pytest: 2; eval: 2 | pr, main, release |
world.fork.lineage |
world |
high | docs/guide/world-lifecycle.md — 4. fork_world |
pytest: 2; eval: 1 | pr, main, release |
world.run_identity.cold_resume |
world |
high | docs/guide/world-lifecycle.md — 6. resume_world (fenced mutable cold resume) |
pytest: 2; eval: 1 | pr, main, release |
world.mutations.compose_deterministically |
world |
high | docs/guide/specification.md — Mutation materialization | pytest: 2; eval: 4 | pr, main, release |
world.mutation.locking |
world |
high | docs/guide/specification.md — Mutation Contracts | pytest: 2 | pr, main, release |
query.cold_reads.correct |
world |
high | docs/guide/specification.md — Durable world reads | pytest: 3; eval: 3; benchmark: 1 | pr, main, release |
storage.cache.concurrent_no_loss |
storage |
high | docs/guide/stores.md — Write-Behind Cache | pytest: 1; eval: 2; benchmark: 1 | pr, main, release |
ingestion.envelope.append_selection |
artifacts |
high | docs/guide/artifacts.md — 5. Catalog and index contract | pytest: 1; static: 2 | pr, main, release |
ingestion.catalog.cold_roundtrip |
artifacts |
high | docs/guide/artifacts.md — 5. Catalog and index contract | pytest: 2 | pr, main, release |
artifacts.ingestion.occurrence_identity |
artifacts |
high | docs/guide/artifacts.md — 4. Occurrence and content identity | pytest: 1; eval: 1 | pr, main, release |
artifacts.ingestion.common_visibility |
artifacts |
high | docs/guide/artifacts.md — 7. Visibility and failure | pytest: 1; static: 1 | pr, main, release |
artifacts.context.task_anchored |
artifacts |
medium | docs/guide/artifacts.md — 10. Task-anchored artifact context | pytest: 1; static: 2 | pr, main, release |
evaluation.result.snapshot_pinned |
evaluation |
high | docs/guide/artifacts.md — 8. Evaluation results | pytest: 2; eval: 1 | pr, main, release |
missions.transcripts.redacted_ingestion |
missions |
high | docs/missions/transcripts.md — Transcript ingestion contract | pytest: 2 | pr, main, release |
missions.transcripts.fail_closed |
missions |
high | docs/missions/transcripts.md — Transcript ingestion contract | pytest: 1 | pr, main, release |
missions.transcripts.occurrence_identity |
missions |
medium | docs/missions/transcripts.md — Transcript ingestion contract | pytest: 1 | pr, main, release |
evaluation.ontology.separated |
evaluation |
medium | docs/guide/dataset-eval-ontology.md — 1. The contract in one view | pytest: 1; eval: 2 | pr, main, release |
audit.append_only.evidence |
commands |
high | docs/guide/audit-log.md — 2. Append-only invariant | pytest: 1; eval: 1 | pr, main, release |
simulation.episode.termination |
world |
medium | docs/guide/specification.md — Run contract | pytest: 1; eval: 3 | pr, main, release |
physical_ai.workflow.evidence |
physical_ai |
high | docs/guide/physical-ai.md — Committed-state sequence | pytest: 4 | pr, main, release |
physical_ai.activity.recovery |
physical_ai |
high | docs/guide/physical-ai.md — Hosted-episode recovery | pytest: 2 | pr, main, release |
core.ecs.data_model |
core |
high | docs/guide/specification.md — Data Model Contracts | pytest: 2; eval: 2; benchmark: 1 | pr, main, release |
core.processors.execution |
core |
high | docs/guide/specification.md — System and Processor Contracts | pytest: 3; eval: 1 | pr, main, release |
core.hooks.lifecycle |
core |
medium | docs/guide/specification.md — Lifecycle Hook Contracts | pytest: 3 | pr, main, release |
missions.trajectory.runtime_service |
missions |
medium | docs/guide/trajectories.md — Select one evidence table | pytest: 3 | pr, main, release |
research.autoresearch.ledger |
research |
high | docs/guide/autoresearch.md — The Loop | pytest: 8 | pr, main, release |
graph.relations.temporal |
graph |
medium | docs/guide/prefab-libraries.md — Current relation and temporal-view contract | pytest: 4 | pr, main, release |
graph.prefabs.copy_on_instantiate |
graph |
medium | docs/guide/prefab-libraries.md — Instantiation is a ledger operation | pytest: 4 | pr, main, release |
missions.sandbox.checkpoint_restore |
missions |
high | docs/guide/agent-missions.md — Supported sandbox backends | pytest: 3 | pr, main, release |
release.package.installable |
release |
high | docs/guide/api-stability.md — What counts as public | pytest: 2; static: 1 | release |
storage.remote.control_parity |
storage |
high | docs/guide/application-architecture.md — 10. Supported durability profile | pytest: 2; static: 1 | main, release |
missions.mcp.trusted_host_adapter |
missions |
high | docs/guide/agent-missions.md — 11. Mission MCP server | pytest: 3 | pr, main, release |
missions.agent_v1.public_authoring |
missions |
medium | docs/guide/agent-missions.md — 2. Public authoring surface | pytest: 2 | pr, main, release |
missions.agent_v1.validator_gated |
missions |
high | docs/guide/agent-missions.md — 4. State and transition protocol | pytest: 7; static: 2 | pr, main, release |
missions.agent_v1.exact_head_critic |
missions |
high | docs/guide/agent-missions.md — Independent exact-head critic | pytest: 4 | pr, main, release |
missions.sandbox.cleanup_ownership |
missions |
high | docs/guide/agent-missions.md — 5. Sandbox and validator protocol | pytest: 2 | pr, main, release |
missions.control.authenticated_profiles |
missions |
high | docs/guide/agent-missions.md — Authenticated execution profiles | pytest: 3; static: 1 | pr, main, release |
security.redaction.pre_durability |
redaction |
high | docs/guide/artifacts.md — 9. Security boundary | pytest: 2; eval: 1 | pr, main, release |
Profile membership states where a contract must be enforced. Eval suite
blocking/advisory policy is defined separately in quality/eval_profiles.toml;
benchmark policy is defined in quality/benchmarks.toml.