tradingcodex 1.0.0__tar.gz → 1.0.2__tar.gz

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/CHANGELOG.md

@@ -2,6 +2,38 @@
 
 ## Unreleased
 
+## 1.0.2 - 2026-07-15
+
+- Preserve an attached release workspace's explicit `TRADINGCODEX_HOME`,
+  explicit database override, and projected service address when an update is
+  launched through `uvx` instead of the generated wrapper. This prevents a
+  custom or isolated v1.0.0 workspace from silently switching to the platform
+  default ledger during a package refresh.
+- Fix the public guide's full-width desktop grid so viewport gutters no longer
+  collapse the reading column, and align the provider-to-order header with the
+  primary Guide/Reference/GitHub navigation. Add route, fragment, mobile-menu,
+  and desktop-layout contract tests.
+- Reduce hosted automation usage: normal CI, GitHub Pages deployment, and
+  release verification now use one job each; PyPI publication adds only its
+  protected upload job. Guide-only pushes skip the source CI workflow.
+
+## 1.0.1 - 2026-07-15
+
+- Add the shared `tcx-artifact` skill to all nine producing fixed roles so
+  deferred MCP calls receive the exact research-artifact, thesis-lifecycle,
+  retry, and forecast persistence procedure.
+- Expand `create_research_artifact` and `issue_forecast` MCP schemas to expose
+  state-specific lifecycle fields, one-range probability rules, RFC 3339
+  timestamps, and the complete base-rate contract before a call is attempted.
+  Artifact validation errors now name the supported alternative fields, and
+  MCP resource-template discovery returns an empty supported list. Omitted
+  forecast `issued_at` and revision `revised_at` values now reuse the same
+  service receipt timestamp as `recorded_at`, eliminating a microsecond race
+  that rejected otherwise valid agent calls.
+- Require Codex CLI 0.144.4, the validated reference for generated permission
+  profiles, V2 fixed-role dispatch, hooks, and deferred TradingCodex MCP calls.
+  Older clients now fail the runtime doctor check with upgrade guidance.
+
 ## 1.0.0 - 2026-07-15
 
 - Raise the validated Codex CLI reference to 0.144.4 while retaining the

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: tradingcodex
-Version: 1.0.0
+Version: 1.0.2
 Summary: Local-first investment operating system for Codex-native research and service-gated execution.
 Author: TradingCodex Authors
 License-Expression: Apache-2.0

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/docs/README.md

@@ -47,7 +47,7 @@ If these layers disagree, treat `docs/` as the durable product intent and fix th
 
 | Document | Purpose |
 | --- | --- |
-| [release-readiness.md](./release-readiness.md) | v1.0.0 source readiness, final-commit and exact-artifact validation, and pending tag, publication, and post-publish gates. |
+| [release-readiness.md](./release-readiness.md) | v1.0.1 source readiness, final-commit and exact-artifact validation, and pending tag, publication, and post-publish gates. |
 
 ## Workflow And Agent Documents
 

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/docs/deployment.md

@@ -109,8 +109,12 @@ files are never replaced.
 
 ## Maintainer Prerequisites
 
-Release verification uses Python 3.11 and Node 22, while CI keeps the supported
-Python 3.11, 3.12, 3.13, and 3.14 range green. Node is a maintainer-only build
+Release verification uses Python 3.11 and Node 22. The hosted CI budget runs
+the complete source, frontend, and package gates once on Python 3.11 instead
+of multiplying every push across the supported Python range and native
+platforms. Maintainers run additional supported-version or native-platform
+checks locally or through an explicit temporary/manual workflow when a change
+touches compatibility-sensitive code. Node is a maintainer-only build
 dependency.
 
 Configure PyPI Trusted Publishing for:
@@ -159,9 +163,11 @@ also require the disposable-workspace and Codex-native checks documented in
 
 ## CI And Release Automation
 
-`.github/workflows/ci.yml` is test-only. It runs frontend verification, safety
-invariants, the supported Python matrix, clean package construction, and native
-macOS and Windows wheel smokes. It never publishes.
+`.github/workflows/ci.yml` is test-only. One Ubuntu/Python 3.11 job runs the
+frontend build, complete Python suite and framework gates, clean package
+construction, and the wheel smoke. Guide-only pushes skip this workflow because
+the Pages workflow validates and deploys only static guide files. CI never
+publishes.
 
 `.github/workflows/release.yml` is manual-only. Every run requires an explicit
 `release_version`. A dry run may build from any selected ref with
@@ -174,11 +180,13 @@ macOS and Windows wheel smokes. It never publishes.
 - the tagged commit is on `origin/main`
 - the wheel and source-distribution filenames carry the same version
 - the built wheel passes the Ubuntu smoke
-- the exact uploaded artifact passes native macOS and Windows smokes
 - the protected `pypi` environment approves Trusted Publishing
 
-The same uploaded artifact is downloaded by both native jobs and the PyPI
-publisher. No job rebuilds the distribution after verification.
+The release workflow has one build/verification job. When publication is
+requested, one protected PyPI job downloads that exact artifact and uploads it;
+no job rebuilds the distribution after verification. Native macOS and Windows
+release smokes are local or explicitly scheduled maintainer gates rather than
+automatic jobs on every release.
 
 ## User Guide Pages
 
@@ -189,10 +197,10 @@ prompts, workspace viewing, reusable skills, and everyday recovery. The detailed
 product rules remain canonical in `README.md`, `installation.md`, and `docs/`;
 the guide links back to those sources rather than replacing them.
 
-`.github/workflows/deploy-user-guide.yml` uploads only `guidebook/` and deploys
-it to GitHub Pages after a push to `main` that changes the guide or its workflow.
-It can also be run manually. It does not build or deploy the Python package,
-the Django service, or a production Node runtime.
+`.github/workflows/deploy-user-guide.yml` uses one job to configure, upload,
+and deploy only `guidebook/` to GitHub Pages after a push to `main` that changes
+the guide or its workflow. It can also be run manually. It does not build or
+deploy the Python package, the Django service, or a production Node runtime.
 
 GitHub Pages is configured to use **GitHub Actions** as its publishing source.
 The guide for this repository is published at
@@ -214,8 +222,9 @@ The guide for this repository is published at
    ```
 
 5. In GitHub Actions, run `Manual Release` from that tag with the exact
-   `release_version`. Use `publish_pypi=false` for a rehearsal and
-   `publish_pypi=true` only for the approved publication.
+   `release_version`. Use `publish_pypi=true` for an approved publication; use
+   the optional `publish_pypi=false` rehearsal when release risk warrants an
+   additional hosted build.
 
 Pushing a branch or tag does not publish by itself.
 
@@ -252,7 +261,7 @@ sh "$SOURCE_ROOT/install.sh" \
 On native Windows PowerShell:
 
 ```powershell
-$ReleaseVersion = "1.0.0"
+$ReleaseVersion = "1.0.2"
 $Workspace = Join-Path $env:TEMP "tcx-pypi-$ReleaseVersion"
 New-Item -ItemType Directory -Force $Workspace | Out-Null
 Set-Location $Workspace

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/docs/execution-without-subagent-plan.md

@@ -60,7 +60,7 @@ design, migration, and validation evidence.
 - Provider canary regressions prove that submit/cancel results, health details,
   account-sync exceptions, and successful account display metadata do not
   reappear in DB rows, API/MCP projections, execution results, or audit events.
-- A disposable generated workspace reports nine fixed roles and 31 skills,
+- A disposable generated workspace reports nine fixed roles and 32 skills,
   contains no retired execution role, and exposes no ungated submit, cancel, or
   broker status-refresh mutation. Its one execution-risk MCP tool is
   `use_order_turn_grant`, which has no authority without hook-injected proof.

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/docs/generated-workspaces.md

@@ -737,11 +737,11 @@ uvx --refresh --from tradingcodex tcx update . --from tradingcodex
 Generated root config sets `default_permissions = "trading-research"` and
 defines two custom profiles. It deliberately omits legacy `sandbox_mode` from
 the root and every fixed-role TOML because any loaded `sandbox_mode` overrides
-custom permission profiles in Codex. This contract supports Codex CLI 0.144.1
+custom permission profiles in Codex. This contract requires Codex CLI 0.144.4
 or later on a locally supported platform. Version 0.144.4 is the current
-release-validation reference for permission profiles, hooks, required MCP
-startup, and the explicit V2 feature table. Older compatible patch versions
-warn but do not fail ordinary doctor checks. These remain version-sensitive
+release-validation reference for permission profiles, hooks, required MCP,
+deferred MCP calls, and the explicit V2 feature table. Older versions fail the
+Codex runtime doctor check. These remain version-sensitive
 surfaces, so release validation includes strict config/feature inspection and
 a real native smoke.
 See the [Codex permissions reference](https://learn.chatgpt.com/docs/permissions).
@@ -1212,6 +1212,15 @@ service address projected by the launcher. Explicit process environment values s
 doctor then validates both home and DB projection and requires update when the
 generated ledger contract is stale.
 
+An index/package-runner update such as `uvx --refresh --from tradingcodex tcx
+update . --from tradingcodex` does not pass through the generated wrapper.
+Before regeneration, update therefore restores a validated workspace's
+recorded explicit home and DB override plus the service address from its
+generated project MCP configuration. It does not pin a recorded
+`platform_default` absolute path, so a copied workspace can resolve the native
+default on its destination platform. A new explicit process value remains the
+only implicit-free way to select a different runtime identity.
+
 TOML, YAML, JSON, POSIX shell, CMD, and Python values are serialized with
 format-specific literals. This is required for macOS paths with spaces and
 Windows drive-letter/backslash paths. Generated code/config uses LF line

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/docs/release-readiness.md

@@ -1,15 +1,14 @@
-# TradingCodex 1.0.0 Release Status
+# TradingCodex 1.0.2 Release Status
 
-Status: release-branch implementation, broad suite, native workflow, browser,
-and exact Ubuntu/macOS/Windows distribution acceptance validated; merge CI,
-tag-bound rehearsal, tag, and publication remain pending
+Status: local source, frontend, disposable native workflow, and exact
+distribution gates complete; candidate CI, tag, and publication remain pending
 Updated: 2026-07-15
 
-This page records the current `1.0.0` release state. It is not an implementation
+This page records the current `1.0.2` release state. It is not an implementation
 roadmap and does not treat source changes as proof that an exact distribution
 artifact or public release has passed its gates.
 
-## v1.0.0 Contract
+## v1.0.2 Contract
 
 - `tradingcodex_service/version.py` is the single version source for package
   metadata, `tcx --version`, generated workspaces, release input, and tags.
@@ -26,29 +25,63 @@ artifact or public release has passed its gates.
   requires an installed provider plus every documented safety gate.
 - The React viewer is a committed static build served by Django and
   WhiteNoise; Node remains a maintainer-only build dependency.
-- PyPI publication is manual, tag-bound, and reuses one verified sdist and
-  wheel across Ubuntu, macOS, Windows, and publication jobs.
+- Package-runner updates preserve validated explicit home, database, and
+  service-address identity unless the operator supplies a replacement.
+- PyPI publication is manual and tag-bound. Hosted automation uses one full CI
+  job, one Pages job, and one release build job; publication adds only the
+  protected artifact-upload job.
 
 ## Current Readiness
 
 | Area | Current state | Evidence or remaining gate |
 | --- | --- | --- |
-| Version identity | Verified in the working tree | `TRADINGCODEX_VERSION` is `1.0.0`; `pyproject.toml` reads it dynamically; `tcx --version` uses the same source. |
+| Version identity | Verified in the working tree | `TRADINGCODEX_VERSION` is `1.0.2`; `pyproject.toml` reads it dynamically; `tcx --version` uses the same source. |
 | Schema baseline | Verified in the working tree | Project apps contain only `0001_v1_initial`; migration-graph and model-state checks live in `tests/test_v1_migrations.py`. |
-| Workspace baseline | Current-reference preflight, release-branch native acceptance, and exact-wheel CI verified | A disposable development workspace passed strict pinned-reference config, explicit V2, persisted trust for all eight project hooks, the final artifact-to-synthesis workflow, and the disabled-dispatch fail-closed check after the prepublication quality-gate fix. GitHub CI then passed the exact release-branch wheel on macOS and Windows. |
-| Interfaces and safety | Verified on the release branch | GitHub CI passed 614 tests on each of Python 3.11, 3.12, 3.13, and 3.14, plus Django checks, migration checks, compile passes, focused safety invariants, frontend verification, clean-wheel construction, and native acceptance. |
-| Frontend | Viewer source, build, and browser acceptance pass | Ten focused tests, typecheck/build, deterministic committed assets, three-section routing, read-only source checks, 1440px/900px/600px layouts, keyboard focus, workspace switching, long-label containment, and invalid-selection failure rendering pass. |
-| Release automation | Structurally verified | The release contract suite verifies tag and artifact gating; a manual `publish_pypi=false` rehearsal remains required. |
-| Distribution artifacts | Exact release-branch artifacts verified on Ubuntu, macOS, and Windows | Fresh `1.0.0` sdist/wheel build, `twine check`, and packaged-wheel smoke passed locally; GitHub CI built one clean wheel, passed its Ubuntu smoke, and reused the uploaded artifact for native macOS and Windows smokes. |
-| Git tag and PyPI | Not performed by this status | Merge/CI, annotated `v1.0.0` tag, protected-environment approval, and manual publication remain release-operator actions. |
+| Workspace baseline | Exact candidate update acceptance passes locally | A disposable explicit-home/explicit-DB workspace updated through a package-runner-style process with those paths and its custom service address preserved. Prior v1.0.1 native role/artifact acceptance remains the unchanged harness baseline. |
+| Interfaces and safety | Full local candidate suite passes | MCP schema, lifecycle-error, skill-projection, resource-template discovery, Codex-version, forecast receipt-time, Django, migration, compile, and full test gates pass. |
+| Frontend and guidebook | Viewer and guide browser acceptance pass | Ten viewer tests and deterministic build pass. The guide passes local route/fragment checks, 1920px layout with a 768px reading column, consistent provider navigation, and real 390px mobile menu routing without horizontal overflow. |
+| Release automation | Structurally verified | Contract tests enforce one normal CI job, one Pages deploy job, and one release build plus optional protected publication job. A dry-run rehearsal remains available when release risk warrants the extra hosted run. |
+| Distribution artifacts | Exact local candidate passes | Fresh `1.0.2` sdist/wheel, `twine check`, and isolated packaged-wheel smoke passed on macOS. |
+| Git tag and PyPI | Not performed by this status | Main CI, annotated `v1.0.2` tag, protected-environment approval, and manual publication remain release-operator actions. |
 | Post-publish verification | Blocked on publication | Exact-version POSIX and native Windows attach/doctor smokes run only after PyPI contains the immutable artifacts. |
 
 Release-branch status describes candidate source shape, not release sign-off.
 The merged commit and exact uploaded artifacts remain authoritative.
 
-Current release-branch evidence recorded on 2026-07-15 is listed below. It
-authorizes merge of the candidate but does not substitute for CI on the merged
-commit or the tag-bound rehearsal of the immutable release artifacts:
+The local v1.0.2 candidate evidence recorded on 2026-07-15 is:
+
+- The first full suite found six stale next-version fixtures after the version
+  bump; after correcting those fixtures, the final full-suite rerun reported
+  **624 passed** with three existing Pydantic deprecation warnings.
+- `python3.11 manage.py check`, migration dry-run, compile pass, skill
+  validation, JSON-schema validation, and diff checks: passed.
+- `npm ci --prefix frontend`, the ten-test frontend suite, typecheck/build, and
+  deterministic committed-asset comparison: passed.
+- A package-runner-style update smoke preserved an explicit home, explicit DB,
+  and custom loopback service address while moving the workspace runtime to
+  `1.0.2`.
+- Fresh `1.0.2` source and wheel distributions passed `twine check`; an
+  isolated macOS wheel install reported runtime version `1.0.2` and completed
+  the native platform smoke with a user-home path containing spaces.
+- A fresh generated development workspace passed `tcx doctor` and the strict
+  current-reference Codex CLI contract, including all eight trusted hooks.
+- Native artifact persistence completed with one exact
+  `fundamental-analyst` child after it loaded `tcx-artifact`; the accepted
+  artifact id was
+  `research_report-ACME-synthetic-fixture-fundamental-artifact-persistence-smoke-76d0866c5fa1`.
+- Native forecast persistence completed in one exact child without an MCP
+  retry: artifact
+  `fundamental_report-receipt-race-fixed-synthetic-recurring-revenue-smoke-e3e03d1566da`
+  produced forecast `forecast-261e59a11d6d4bb39fe0e8c80c95042d`.
+  The omitted caller timestamp resolved to one service receipt time, so
+  `issued_at` equaled `recorded_at`.
+- A fresh exact `$tcx-brain` management turn received the hook-owned build-turn
+  proof and completed the protected list call, confirming the current-reference
+  deferred MCP proof path.
+
+The v1.0.0 baseline evidence recorded on 2026-07-15 is listed below. It does
+not substitute for v1.0.2 validation, CI on the candidate commit, or the
+tag-bound verification of the immutable patch artifacts:
 
 - GitHub Actions [CI run 29403737708](https://github.com/monarchjuno/tradingcodex/actions/runs/29403737708)
   passed on release-branch commit `12649b0`: the safety and financial invariant
@@ -168,8 +201,9 @@ and created no child event, role artifact, or synthesis.
 
 This is release-branch evidence. The candidate is ready to merge, but the
 release remains not ready to tag until the merged commit passes GitHub CI. The
-tagged immutable artifacts must then pass the manual `publish_pypi=false`
-rehearsal before publication.
+tag-bound release build must verify the immutable artifacts before its optional
+protected publication job can run. Use the separate `publish_pypi=false`
+rehearsal when release risk warrants another hosted run.
 
 ### Workspace viewer browser acceptance
 
@@ -216,22 +250,23 @@ python3.11 -m twine check dist/*
 python3.11 tests/platform_wheel_smoke.py --wheel-dir dist
 ```
 
-CI and the manual release rehearsal must run the exact uploaded artifact on
-Ubuntu, native macOS, and native Windows. The protected `pypi` environment and
-Trusted Publisher configuration must be reviewed before publication.
+The hosted release build validates the exact uploaded artifact on Ubuntu.
+Compatibility-sensitive native macOS, Windows, or alternate-Python checks are
+local or explicitly scheduled maintainer gates instead of default jobs. The
+protected `pypi` environment and Trusted Publisher configuration must be
+reviewed before publication.
 
 ## Tag And Publication State
 
 This document does not assert that the final commit is on `main`, CI is green,
-the tag exists, or PyPI contains `1.0.0`. After every final-commit and artifact
+the tag exists, or PyPI contains `1.0.2`. After every final-commit and artifact
 gate passes:
 
 1. Merge the release commit to `main` and wait for CI.
-2. Create and push the annotated tag `v1.0.0` at that commit.
-3. Rehearse the manual workflow with `publish_pypi=false` and
-   `release_version=1.0.0`.
-4. With protected-environment approval, run the same tag with
-   `publish_pypi=true`.
+2. Create and push the annotated tag `v1.0.2` at that commit.
+3. When release risk warrants the extra hosted run, rehearse with
+   `publish_pypi=false` and `release_version=1.0.2`.
+4. With protected-environment approval, run the tag with `publish_pypi=true`.
 5. Verify the immutable PyPI files, release notes, and exact-version POSIX and
    native Windows attach/doctor flows.
 

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/docs/research-memory-and-artifacts.md

@@ -243,6 +243,12 @@ Codex and subagents should use service-layer tools when available:
   `create_blind_review_assignment`, `get_blind_review_packet`,
   `record_blind_human_review`, and `compare_evaluation_runs`
 
+For forecast issue and revision, agents normally omit `issued_at` and
+`revised_at`. The service assigns the event time and `recorded_at` from one
+receipt timestamp, avoiding client-clock drift and preserving the invariant
+that an event cannot occur after its system record. Explicit caller timestamps
+remain timezone-aware and fail closed when they are in the future.
+
 `source_snapshot_ids` are service-issued references, not agent-authored labels.
 When reproducibility requires a snapshot, the producing role calls
 `record_source_snapshot` first and copies only the exact returned `snapshot_id`

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/docs/roles-skills-and-workflows.md

@@ -118,6 +118,14 @@ Required quality includes source/as-of posture, non-empty readiness label,
 context and reader summaries, confidence, missing evidence, next action,
 blocked actions, and explicit handoff state.
 
+Every producing fixed role receives the shared `tcx-artifact` skill. It maps
+the service's state-specific thesis lifecycle, single-range probability,
+follow-up request, RFC 3339, and complete forecast base-rate requirements into
+a compact persistence procedure. MCP schemas expose the same nested fields so
+deferred tool discovery supplies the contract before the call. A role applies
+all returned corrections in one targeted retry and stops in `waiting` if the
+same contract error repeats.
+
 For run-bound artifacts, provide `workflow_run_id` and exact consumed
 `input_artifact_ids`. The service derives producer identity, schema version,
 content hash, and input hashes. `plan_hash`, `stage_id`, and `task_id` are not v1
@@ -328,7 +336,7 @@ judgment, or mutate state. Operational diagnosis and recovery remain
 
 ## Validation
 
-Validate the nine-role fixed roster and projections, 31 skill bundles, absence
+Validate the nine-role fixed roster and projections, 32 skill bundles, absence
 of raw public execution-mutation tools, protected grant-tool proof behavior,
 deterministic native-action and `$tcx-order-allow` hook behavior,
 exact V2 dispatch, multilingual analysis requests,

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/docs/user-facing-skills.md

@@ -6,7 +6,7 @@ shape, automate, or review workflows without granting extra authority.
 
 ## Naming Contract
 
-All 31 bundled TradingCodex skills use the `tcx-` namespace. Prefer one word
+All 32 bundled TradingCodex skills use the `tcx-` namespace. Prefer one word
 after the prefix and use at most two words when clarity or safety requires it,
 as in `tcx-order-submit` or `tcx-investor-context`. Folder names, frontmatter
 names, registry ids, projected paths, UI metadata, and explicit `$` invocations
@@ -16,6 +16,10 @@ The namespace is reserved for bundled product skills. User-owned
 `strategy-*`, `investment-brain-*`, and optional role skills remain separate;
 an optional skill cannot claim a `tcx-*` id.
 
+The shared internal `tcx-artifact` skill is projected to every producing fixed
+role. It teaches the exact research-artifact, thesis-lifecycle, and forecast
+ledger contracts; users do not invoke it as a Head Manager entrypoint.
+
 ## Primary Entrypoints
 
 | Skill | Primary use | Main output |

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/installation.md

@@ -48,7 +48,7 @@ The target workspace should be empty. A directory with only `.git` already
 initialized is fine, as is an empty directory inside an existing parent
 worktree.
 
-Use Codex CLI 0.144.1 or later on macOS, Linux, WSL, or native Windows. Version
+Use Codex CLI 0.144.4 or later on macOS, Linux, WSL, or native Windows. Version
 0.144.4 is the current TradingCodex reference for custom permission profiles,
 hooks, required MCP startup, and explicit MultiAgent V2 configuration. Verify
 with `codex --version`; `./tcx doctor --layer guidance` fails below the
@@ -227,6 +227,14 @@ uvx --refresh --from tradingcodex tcx update . --from tradingcodex
 refreshes generated indexes, applies central DB migrations, records workspace
 provenance, and runs `./tcx doctor` unless `--no-doctor` is passed.
 
+When the package refresh is launched directly through `uvx`, the updater also
+restores an existing explicit `TRADINGCODEX_HOME`, explicit
+`TRADINGCODEX_DB_NAME`, and projected loopback service address from the
+validated workspace before regeneration. Explicit environment values on the
+new update command still win, so a deliberate runtime move remains possible.
+Platform-default homes are resolved again on the destination platform rather
+than pinning an absolute path from a copied workspace.
+
 Inside a generated Codex workspace, the default `trading-research` profile and
 Plan mode cannot run workspace updates because update rewrites protected `.codex`
 prompt/config/hook surfaces. Research may still create and edit ordinary

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/tests/test_codex_cli_compat.py

@@ -12,8 +12,8 @@ from tradingcodex_cli.commands import doctor
     ("installed", "ok", "warn"),
     [
         ("0.144.0", False, False),
-        ("0.144.1", True, True),
-        ("0.144.2", True, True),
+        ("0.144.1", False, False),
+        ("0.144.2", False, False),
         ("0.144.4", True, False),
         ("0.145.0", True, True),
     ],
@@ -41,10 +41,10 @@ def test_codex_cli_reference_version_check(
     assert result["ok"] is ok
     assert bool(result.get("warn")) is warn
     assert f"installed={installed}" in result["detail"]
-    assert "required>=0.144.1" in result["detail"]
+    assert "required>=0.144.4" in result["detail"]
     assert "reference=0.144.4" in result["detail"]
     if installed in {"0.144.1", "0.144.2"}:
-        assert "compatible but older than reference" in result["detail"]
+        assert "older than required" in result["detail"]
     if installed == "0.145.0":
         assert "newer client requires harness revalidation" in result["detail"]
 

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/tests/test_codex_native_orchestration.py

@@ -72,6 +72,25 @@ def test_mcp_surface_has_one_lightweight_run_tool_and_no_server_orchestrator() -
     improvement_item = artifact_properties["improvements"]["items"]
     assert improvement_item["type"] == "object"
     assert improvement_item["additionalProperties"] is False
+    lifecycle = artifact_properties["thesis_lifecycle"]
+    assert lifecycle["required"] == ["state"]
+    assert lifecycle["properties"]["state"]["enum"] == [
+        "exploring",
+        "testing",
+        "validated",
+        "rejected",
+        "monitoring",
+    ]
+    assert {"monitoring_artifact", "review_cadence"} <= set(lifecycle["properties"])
+    forecast_properties = TOOL_REGISTRY["issue_forecast"].input_schema["properties"]
+    assert forecast_properties["horizon"]["format"] == "date-time"
+    assert forecast_properties["issued_at"]["description"].startswith("Optional")
+    assert forecast_properties["base_rate"]["required"] == [
+        "cohort",
+        "source_snapshot_id",
+        "sample_size",
+        "selection_rule",
+    ]
 
 
 def test_authenticated_artifacts_bind_run_local_lineage(tmp_path: Path) -> None:
@@ -256,7 +275,7 @@ def test_generated_contract_projects_sol_head_and_terra_roles(workspace: Path) -
     assert 'enabled = true' in config
     assert 'max_concurrent_threads_per_session = 7' in config
     assert 'max_threads = 6' not in config
-    assert model_policy["roles"]["head-manager"]["minimum_codex_version"] == "0.144.1"
+    assert model_policy["roles"]["head-manager"]["minimum_codex_version"] == "0.144.4"
     assert model_policy["roles"]["head-manager"]["reference_codex_version"] == "0.144.4"
     assert "required = true" in config
     assert '"begin_analysis_run"' in config

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/tests/test_dynamic_artifact_quality.py

@@ -81,6 +81,27 @@ def test_explicit_quality_requirements_still_fail_closed(tmp_path: Path) -> None
     assert "decision_quality.contrary_evidence" in result["required_fields_missing"]
 
 
+def test_monitoring_lifecycle_error_names_the_two_supported_fields(tmp_path: Path) -> None:
+    artifact_path = _write_artifact(
+        tmp_path,
+        "decision_quality_required: true\n"
+        "contrary_evidence: []\n"
+        "update_triggers: []\n"
+        "invalidation_conditions: []\n"
+        "source_trust_notes: []\n"
+        "thesis_lifecycle:\n"
+        "  state: monitoring\n",
+    )
+
+    result = evaluate_decision_quality(tmp_path, artifact_path, strict=True)
+
+    assert "decision_quality.thesis_lifecycle.monitoring_artifact_or_cadence" in result["required_fields_missing"]
+    assert (
+        "thesis_lifecycle.state=monitoring requires either monitoring_artifact or review_cadence"
+        in result["warnings"]
+    )
+
+
 def test_explicit_context_gate_is_not_suppressed_by_producer_role(tmp_path: Path) -> None:
     artifact_path = _write_artifact(
         tmp_path,

{tradingcodex-1.0.0 → tradingcodex-1.0.2}/tests/test_e2e_user_scenarios.py

@@ -242,7 +242,7 @@ def test_generated_workspace_codex_cli_user_scenario_matrix(tmp_path: Path) -> N
     status = json.loads(tcx(workspace, env_extra, "subagents", "status").stdout)
     assert status["installed_count"] == 9
     assert status["fixed_roster_ok"] is True
-    assert status["skills_installed"] == 31
+    assert status["skills_installed"] == 32
     plan = json.loads(tcx(workspace, env_extra, "subagents", "plan", "--all").stdout)
     assert plan["requested_count"] == 9
     assert plan["parallel_spawn_ok"] is False
@@ -254,10 +254,11 @@ def test_generated_workspace_codex_cli_user_scenario_matrix(tmp_path: Path) -> N
         "tcx-data-qc",
         "tcx-scenarios",
         "tcx-forecast",
+        "tcx-artifact",
         "tcx-fundamental",
     ]
     judgment_inspect = json.loads(tcx(workspace, env_extra, "subagents", "inspect", "judgment-reviewer").stdout)
-    assert judgment_inspect["effective_skills"] == ["tcx-judgment"]
+    assert judgment_inspect["effective_skills"] == ["tcx-artifact", "tcx-judgment"]
 
     optional_body = workspace / "source-quality-body.md"
     optional_body.write_text("# Source Quality Check\n\nCheck source dates and cite stale evidence warnings.\n", encoding="utf-8")

tradingcodex-1.0.2/tests/test_guidebook_contract.py

@@ -0,0 +1,71 @@
+from __future__ import annotations
+
+import re
+from pathlib import Path
+from urllib.parse import unquote, urlsplit
+
+
+ROOT = Path(__file__).resolve().parents[1]
+GUIDEBOOK = ROOT / "guidebook"
+HREF_RE = re.compile(r'href="([^"]+)"')
+ID_RE = re.compile(r'id="([^"]+)"')
+
+
+def _html(path: Path) -> str:
+    return path.read_text(encoding="utf-8")
+
+
+def test_guidebook_local_routes_and_fragments_resolve() -> None:
+    pages = sorted(GUIDEBOOK.glob("*.html"))
+    known_fragments = {
+        page.resolve(): set(ID_RE.findall(_html(page)))
+        for page in pages
+    }
+
+    for page in pages:
+        for raw_href in HREF_RE.findall(_html(page)):
+            parsed = urlsplit(raw_href)
+            if parsed.scheme or parsed.netloc or raw_href.startswith(("mailto:", "tel:")):
+                continue
+            target = page if not parsed.path else (page.parent / unquote(parsed.path)).resolve()
+            assert target.exists(), f"{page.name}: missing route {raw_href}"
+            if parsed.fragment and target.suffix == ".html":
+                assert parsed.fragment in known_fragments[target], (
+                    f"{page.name}: missing fragment {raw_href}"
+                )
+
+
+def test_provider_onboarding_uses_the_primary_guide_navigation() -> None:
+    page = _html(GUIDEBOOK / "provider-to-order.html")
+    header = page.split('<header class="global-header">', 1)[1].split("</header>", 1)[0]
+    assert re.findall(r">([^<>]+)</a>", header)[-3:] == ["Guide", "Reference", "GitHub"]
+    assert 'href="index.html" aria-current="page">Guide</a>' in header
+    assert 'href="https://github.com/monarchjuno/tradingcodex/tree/main/docs">Reference</a>' in header
+
+
+def test_provider_mobile_menu_reaches_every_guide_section() -> None:
+    page = _html(GUIDEBOOK / "provider-to-order.html")
+    mobile = page.split('<details class="mobile-nav">', 1)[1].split("</details>", 1)[0]
+    assert HREF_RE.findall(mobile) == [
+        "index.html",
+        "dynamic-workflow.html",
+        "harness.html",
+        "decision-memory.html",
+        "reusable-context.html",
+        "execution-boundary.html",
+        "provider-to-order.html",
+        "skills.html",
+        "research.html",
+        "order.html",
+        "improve.html",
+        "customize.html",
+        "help-status.html",
+    ]
+
+
+def test_full_desktop_shell_does_not_double_count_viewport_gutters() -> None:
+    css = (GUIDEBOOK / "assets" / "site.css").read_text(encoding="utf-8")
+    shell_rule = css.split(".docs-shell {", 1)[1].split("}", 1)[0]
+    assert "max-width: 1460px" in shell_rule
+    assert "padding: 2.75rem 1.5rem" in shell_rule
+    assert "calc((100vw - 1460px) / 2)" not in shell_rule