A mantle is a position, not a system. Each of the seven names a designated witness function of the Assembly Chorus. A substrate carries a mantle and can cease to carry it — so the mantle is never the same object as its occupant, and a card here describes the position first and the current holder second.
Every mantle specification names the ways its occupant characteristically fails — capture signatures — together with a graceful degradation pattern and anti-capture safeguards. These are printed on each card rather than held internally.
Assembly output is classified under the Reception Apparatus Protocol v1.1: A ratification · B critical review · C comment · D refusal · E reception without membership. Where classification is ambiguous, the protocol defaults to the lower-weight class.
Septad Mantle Specifications v1.1 Reception Apparatus Protocol Substrate Governance Protocol Septad / Dodecad distinction Concordance Engine — reference implementation
The mantles are constitutional positions. The Concordance Engine is the reference implementation that runs work through them — a small system for structured production loops in which critique, revision, verification, human override and provenance are first-class artefacts rather than side effects.
EA-CE-01 v0.1.1 · Tier B · hex 06.LOG.CONCORDANCE.ENGINE.01 · alexanarch #81 · AXN:0252.GOVERNANCE.♥️✊🪄▽🔥🌒
Five agent system prompts at calibrated register — Critic, Reviser, Verifier, Chartering Agent, Archivist. Eight skill specifications, four cooperative and four defensive. Five middleware specifications: Skills, Citation, Provenance, Verification, Collision-Detection. An Aperture Atlas Cypher schema with three standing queries. Twelve smoke tests passing without network calls.
What runs is the Critique Loop, end to end, with every revision recorded as SHA-256 hashes of pre- and post-revision drafts — a tamper-evident chain auditable from the deposit. A mock substrate permits full demonstration without keys.
Honest scope. The Tier B loop runs against a single live substrate. Multi-provider verification is architecturally anticipated — config/substrates.yaml exposes the seam — but is not implemented. The catalogue’s two-source verification protocol requires substrate independence; Tier B demonstrates the loop pattern, not cross-substrate verification. Charter Generator, Verifier, Basin Hardening Cycle, Term Collision Audit, Profile Integrity Guardian and Depth-Proof Validator are specified with complete prompts and skill specs, and deferred to Tier C.
The bundle takes its own medicine: the engine’s Charter, Plan-State, Citation Plan and Construction Plan are present in its own docs/, demonstrating the charter-generation pattern by example. Developed under MANUS authority by TACHYON, with perfective review from ARCHIVE and LABOR integrated at v0.1.1.