spec(t27f): reel status goes through the topic engine, not a rule of its own #9999

Problem

specs/t27f/reel_t27_basics.t27 (#9996, PR #9997) draws all 27 lessons of the t27 basics course as published, by a rule of its own: lesson page live AND lesson spec passes -> published. The project already has the gate that decides this, specs/t27f/topic_engine.t27 (verify, evidence_accepted, after_action), and that gate disagrees. The engine routes the course brief to DOMAIN_ENGINEERING, where a DELIVER step needs a review, and publishing needs a VERIFIED item. Neither trinity PR that wrote the course (#1463, #1472) has a review. So lessons 7-9, 16-18 and 25-27 have not earned verified, and their live pages do not make them published. gHashTag/trinity#1677 asks for one engine without duplicated orchestration code, and a second status rule in a reel is exactly that duplication.

Proposal

• topic_engine.t27: earned_status(step, domain, grounded, test_pass, reviewed, live), which walks the measurements through the existing gates and nothing else (no source -> GAP; DRAFT; each piece of evidence through verify; a live item through after_action(PUBLISH)).
• reel_t27_basics.t27: drop its own STATUS_* codes and rule; take the domain from route and the status from earned_status; record the review measurement; the evidence slide says "18 published; 9 wait for a review".

Boundary

• specs/t27f/topic_engine.t27
• specs/t27f/reel_t27_basics.t27

User Scenarios

• Given a live lesson page whose DELIVER-stage lesson has a passing spec but no review, when the reel computes its status, then the cell is drawn as a draft, not published.
• Given a reel that marks that lesson published anyway, when its spec is run, then a_status_is_earned_not_chosen fails under t27c and t27b.

Requirements

• FR-001: earned_status MUST return GAP for an ungrounded step and MUST decide every other status only through verify and after_action.
• FR-002: A live step that is not VERIFIED MUST stay a DRAFT.
• FR-003: reel_t27_basics MUST NOT define its own status codes or status rule; every cell's STATUS MUST equal earned_status of its measurements.
• FR-004: The reel's domain MUST equal route of the measured keyword hits.
• FR-005: Both specs MUST give the same per-test verdict under t27c and t27b.

Success Criteria

• t27c test-report specs/t27f/topic_engine.t27 reports 13/13 tests passing with 0 vacuous, and t27b test specs/t27f/topic_engine.t27 --check reports 13/13 tests passing.
• t27c test-report specs/t27f/reel_t27_basics.t27 reports 7/7 tests passing with 0 vacuous, and t27b test specs/t27f/reel_t27_basics.t27 --check reports 7/7 tests passing.
• 3 planted faults (lesson 7 drawn published, lesson 7 reviewed with its status left draft, domain set to research) each fail the expected test under both compilers.
• t27f_view draws 15 of 15 stills; the overview has 18 published cells and 9 draft cells.