Reading the badgesVerifiedImplementedSyntheticMockPlannedUnverified
- Verified
A named test in the repository exercises exactly this, and the claim cites it. The tests were read and counted for this portal, not executed by it.
- Implemented
The code exists and is documented, but no test, log or report in the repository exercises this specific behaviour.
- Synthetic
Shown only by deterministic experiments over an in-process synthetic lab or a synthetic benchmark — never against a real environment.
- Mock
Exercised only through mock adapters or stubs: no real model provider, no network, no credentials.
- Planned
Described in the dossier, the README or the docs as intended. Not built.
- Unverified
Stated in the repository — a ledger, the README, a commit message — with no test, log or report in it that backs the statement.
06 / 11 · Milestones
Implementation milestones
M1 to M12, each with its commits, its test files and the status the evidence supports.
At a glance
A milestone’s status is the status of what it delivered. A thoroughly tested router that only ever talks to a scripted adapter is still mock; a mission kernel that only ever runs a synthetic lab is still synthetic.
| Milestone | Status | Test functions | Commits |
|---|---|---|---|
| M1 Substrate and audit | Verified | 72 | c260deb 87eb97d a014713 ce581d7 3ecca7d 1d662a6 |
| M2 Permission Engine | Verified | 38 | dd77417 2afb5d9 1d662a6 |
| M3 World Model and Capability Graph | Verified | 90 | 65acbfb 6bfa6b1 |
| M4 Bond Engine and Verification | Verified | 74 | 0effd02 f609d9d f5e7cc4 |
| M5 Persistent memory and evidence projections | Verified | 23 | 6210943 |
| M6 Model abstraction and router | Mock | 21 | 9e813b2 |
| M7 Parasite framework | Verified | 21 | f2b03ed |
| M8 Mission Kernel and Mode Controller | Synthetic | 12 | 7716181 |
| M9–M10 Generations, evaluation and gated self-evolution | Synthetic | 26 | 3b57101 |
| M11–M12 Symba operator interface and end-to-end experiments | Synthetic | 18 | 689eb3e |
Each milestone
- Verified
M1 — Substrate and audit. Core primitives, the append-only hash-chained audit log with replay, the SymbiOS local substrate with its event bus and kill switch, and the local CLI.
Review: Done. Tests: test_core.py, test_audit.py, test_substrate.py, test_cli.py.
Sourcedocs/BUILD_STATUS.md · docs/M1-M2.md @ 9243204 · commits c260deb … 1d662a6
- Verified
M2 — Permission Engine. A pure, audited ALLOW / ASK / DENY engine over TOML policy data, taint-aware and fail-closed, whose every decision replay can re-derive.
Review: Done (with M1). Tests: test_policy_engine.py, test_policy_loader.py, test_policy_matrix.py, test_architecture.py.
Sourcedocs/BUILD_STATUS.md · docs/M1-M2.md @ 9243204 · commits dd77417, 2afb5d9
- Verified
M3 — World Model and Capability Graph. Separated, event-sourced belief and capability state, every mutation authorized by the one Permission Engine and re-checked at replay.
Review: Done and hardened after an adversarial review. Tests: test_world_model.py, test_capability_graph.py, test_domain_common.py, test_domain_authorization.py.
Sourcedocs/BUILD_STATUS.md · docs/M3.md @ 9243204 · commits 65acbfb, 6bfa6b1
- Verified
M4 — Bond Engine and Verification. Deterministic verdicts over pre-registered criteria, and bonds that are earned, decay, and compose only over a fully verified chain.
Review: Done and hardened over two review rounds; 16 surviving mutations still need tests. Tests: test_bond_engine.py, test_bond_hardening.py, test_verification.py, test_verification_hardening.py.
Sourcedocs/BUILD_STATUS.md · docs/M4.md @ 9243204 · commits 0effd02, f609d9d, f5e7cc4
- Verified
M5 — Persistent memory and evidence projections. Evidence-linked research memory with computed provenance, append-only supersession, and operator-gated export and carry.
Review: Done; adversarial review deferred to the final pass. Tests: test_memory.py.
Sourcedocs/BUILD_STATUS.md · docs/M5.md @ 9243204 · commit 6210943
- Mock
M6 — Model abstraction and router. Provider-independent model cards, closed output schemas and an audited, replayable router — exercised only against deterministic scripted adapters.
Review: Done; review deferred. Tests: test_models.py.
Sourcedocs/BUILD_STATUS.md · docs/M6.md @ 9243204 · commit 9e813b2
- Verified
M7 — Parasite framework. Roles as data, a narrow execution context, runtime-enforced footprints and budgets, and a Death Agent audit of every Parasite.
The runtime’s guarantees are tested; the environment it runs is the synthetic lab. Review: Done; review deferred. Tests: test_parasites.py.
Sourcedocs/BUILD_STATUS.md · docs/M7.md @ 9243204 · commit f2b03ed
- Synthetic
M8 — Mission Kernel and Mode Controller. The full mode state machine with CONTROL takeover and External-Controller-only containment, and missions driven through Parasites by a rule-based playbook.
Mode enforcement is tested directly; missions only run against the synthetic lab. Review: Done; review deferred. Tests: test_missions.py.
Sourcedocs/BUILD_STATUS.md · docs/M8.md @ 9243204 · commit 7716181
- Synthetic
M9–M10 — Generations, evaluation and gated self-evolution. Immutable generations with lineage and rollback, sandboxed held-out evaluation, and acceptance that needs a measured improvement and a human approver.
Review: Done; review deferred. Tests: test_evolution.py.
Sourcedocs/BUILD_STATUS.md · docs/M9-M10.md @ 9243204 · commit 3b57101
- Synthetic
M11–M12 — Symba operator interface and end-to-end experiments. A console bound to one principal, an integrity check over the whole log, and six deterministic experiments that run the assembled organism through it.
Review: Done; review deferred. Tests: test_symba.py.
Sourcedocs/BUILD_STATUS.md · docs/M11-M12.md @ 9243204 · commit 689eb3e
What the ledger lists next
- Planned
A final validation pass — adversarial review of M5–M12, mutation-killing tests for M4, and a final report. The ledger lists it as in progress; no report is in the repository.
Sourcedocs/BUILD_STATUS.md (“Final validation pass and report”, “Next actions”) @ 9243204