[VPRamon]

Independent releases currently lack trustworthy producer/consumer evidence. This draft adds exact artifact contract bundles and a deployment-plan gate that checks frontend and MCP independently against Server or Cloud, including unchanged historical consumers. Missing artifacts, mismatched versions/source SHAs/digests, unsupported changed constraints, and incompatible contracts fail with structured diagnostics.

Related to #423. This does not complete the P0 prerequisite and must remain draft.

Implemented

• Inventory of all 245 compiled API operations (241 Server, 245 Cloud), separating origin-root OAuth discovery/protocol routes and excluding the Node MCP transport.
• Explicit schemas for 39 operations: metadata/provider discovery, current user, MCP principal, Cloud account flows, OAuth discovery, and all 25 use-case operations. The remaining 206 operations are explicitly marked unverified; no guessed success schemas are presented as proof.
• Reproducible TypeScript wire models adopted for metadata/principal and selected use-case consumers, generated route expectations, and source-review drift detection.
• Version-controlled MCP compatibility metadata retained in images; unchanged runtime CompatibilityBanner behavior.
• Offline artifact-bundle compatibility command, immutable deployment-plan inputs, PR baseline checks, and reusable CI. Every existing image publisher depends on the checks. The candidate bundle is checked against the exact tagged source, and unchanged artifact bundles are never substituted with main.
• Isolated PostgreSQL integration plus a live Rocket process exercised by the actual frontend transport and MCP client. Recorded traffic excludes credentials.

Validation

• 39 Python positive/negative checks pass, including independent-version fixtures, missing historical artifacts, altered required fields/enums/types/authentication, candidate identity, and genuine process exit codes.
• Compiled Server/Cloud route inventory test passes.
• Real backend smoke passes: 15 audited exchanges, 33 explicitly inventory-only exchanges; corrupted copies of real traffic reject a removed guaranteed response property, wrong response type, and wrong MCP request type.
• Rust workspace tests: 2,061 passed, one ignored; Clippy and rustfmt pass. Default workspace feature unification uses test helpers; the separate isolated production-path run supplies real database/HTTP evidence.
• Frontend: build and 704-test coverage run pass (70.36% lines). One earlier report-builder failure passed in isolation and the full coverage rerun.
• MCP: build and all 51 tests pass. Legacy JS: 411 passed, seven skipped. Stylelint passes; unused-CSS check exits 0 with existing warnings.
• Frontend production npm audit passes. Existing MCP lockfile audit fails with one high/three moderate advisories. Existing Rust advisory audit fails on RUSTSEC-2026-0384, RUSTSEC-2026-0385 and RUSTSEC-2026-0387. No audit gate was relaxed.
• The release completeness check intentionally exits 1 for the remaining 206 operations. Source-bound bundle export was verified after committing.

Merge/release blockers and remaining work

Merging this draft as-is blocks all image publication. Complete the remaining 206 wire audits, per-consumer operation selection and request/response type adoption, broader Cloud/session/OAuth integration coverage, and retention/audit of real historical release contracts. Resolve the existing security failures before release readiness. The current conservative bundles require the entire backend surface and can reject unrelated changes.

The architecture assessment, policy, update procedure, exact changed-file list, commands/results, exclusions and remaining risks are in the HTTP contract report and release documentation. The component-tag architecture is retained; global release migration and deployment remain in #423. Production/VPS configuration was not changed.

[github-actions]

Frontend SPA coverage (by module)

Module	Line Rate	Branch Rate	Function Rate	Health
api	89.1% (311 / 349)	82.7%	86.0%	✔
components	74.1% (2234 / 3015)	62.6%	70.5%	➖
components/graph	100.0% (9 / 9)	66.7%	100.0%	✔
components/reports	86.1% (87 / 101)	70.6%	74.1%	✔
components/table	100.0% (7 / 7)	80.0%	100.0%	✔
context	77.8% (56 / 72)	40.0%	69.6%	➖
hooks	72.3% (146 / 202)	71.1%	70.0%	➖
lib	100.0% (18 / 18)	100.0%	100.0%	✔
pages	68.8% (2332 / 3392)	55.5%	61.1%	➖
pages/catalog	34.8% (162 / 465)	24.5%	18.3%	❌
pages/groups	0.0% (0 / 219)	0.0%	0.0%	❌
src (root)	33.3% (8 / 24)	38.1%	60.0%	❌
types	100.0% (0 / 0)	100.0%	100.0%	✔
utils	97.2% (620 / 638)	89.1%	98.8%	✔
Summary	70.4% (5990 / 8511)	59.1%	65.1%	➖

Generated from Vitest @vitest/coverage-v8 (frontend/src).

[github-actions]

Package	Line Rate	Complexity	Health
marreq-core.src.use_case_document	94%	0	✔
marreq-server.src	57%	0	➖
marreq-core.src.services	93%	0	✔
marreq-core.src.models	80%	0	✔
marreq-core.src.auth	90%	0	✔
marreq-cloud.src.fairings	100%	0	✔
marreq-cloud.src.api	78%	0	✔
marreq-core.src.fairings	90%	0	✔
marreq-core.src.diff	100%	0	✔
marreq-core.src.helper_functions	95%	0	✔
marreq-cloud.src	71%	0	➖
marreq-core.src.api	83%	0	✔
marreq-core.src.auth.rate_limiter	73%	0	➖
marreq-core.src	80%	0	✔
marreq-core.src.importers	74%	0	➖
marreq-core.src.repository.cache	100%	0	✔
marreq-core.src.repository.diesel_repo	5%	0	❌
marreq-core.src.auth.guards	57%	0	➖
marreq-core.src.generators	90%	0	✔
marreq-core.src.routes	61%	0	➖
marreq-core.src.reports	95%	0	✔
marreq-core.src.services.semantic_search	75%	0	➖
marreq-core.src.storage	93%	0	✔
marreq-core.src.deployment	84%	0	✔
marreq-core.src.reqif	97%	0	✔
marreq-cloud.src.services	91%	0	✔
marreq-core.src.validation	96%	0	✔
marreq-core.src.repository	88%	0	✔
Summary	81% (45266 / 56209)	0	✔