📖About

Operator ruling (Ninad, 2026-07-20): the autonomous build pipeline must have NO active LLM. Today the fleet lane is active_provider=kimi and a naive blank-out silently falls back to Claude (Max lane), which violates CAROL-INI-2389. Build a first-class 'disconnected' fleet state: a NullProvider whose call() refuses before spending and never cascades to Claude; make get_active_provider() honor the sentinel and never fall to claude; add llm_switch.py disconnect/reconnect; clear the two Kimi pins on Albus's troubleshooter + RSI diagnosis so they refuse too. Carol's WhatsApp lane (OpenAI, no_fleet_net) and the intake/strategy lane stay live. Scope confirmed by Ninad: fleet only, fail-closed. Filed by the operator via Orion bypass (planner mode is impossible once its own LLM is removed).

⚖️Decisions

• Elrond's bypass methodology checklist (a reminder, not a gate -- you've got this): 0. File it requested_mode='bypass' (planner-vs-bypass is a deliberate choice). bypass_start REFUSES a non-bypass initiative (CAROL-INI-1846), and the dispatcher only skips the bypass lane when the mode says bypass -- a 'planner' mistag lets Merlin's pipeline grab the placeholder step and block your finished work. 1. Filed as planned status -- let the bypass claim/activate it; never file active. 2. Open the bypass (bypass_start) with your droid id + the remediation answer (remediates_initiative_id=NNN, or remediates_nothing=True). 3. Work the blocks for your work-type: template -> design -> code -> test -> review. Do the real work; record decisions on the initiative as you make them. 4. Reality is recorded for you at close -- code (files changed), each decision, and the twin-review verdict become real activities tied to this initiative and show in the Activity Tracker like a planner run (CAROL-INI-1840). No dummy rows. 5. Keep the initiative status moving; it parks in 'reviewing' and is tagged uat-pending for you at close (CAROL-INI-1836), so the stuck-watchdog leaves it alone until UAT. 6. Close runs the gates (design/architecture compliance + caller-audit). If a gate flags something pre-existing or unrelated to your change, waive it with a clear written rationale -- audit, don't skip. 7. Bypass skips the planner's auto-orchestration, NOT the standards. Same template checklist, same review, same observability as a planner run. (elrond)
• [status-router] planned -> executing | event=bypass_executing | bypass transition (or-bx-01)
• [status-router] executing -> reviewing | event=bypass_reviewing | bypass transition (or-bx-01)
• [status-router] reviewing -> closed | event=operator_signoff | Auto-accepted (CAROL-INI-1859): Orion-initiated, >2 days in reviewing with no objection. (el-srac-01)

✅Success criteria

• The provider switch status shows the fleet as 'disconnected' when the pipeline LLM is turned off. (must_have)
• A build-pipeline agent's LLM call returns a clean 'disconnected' refusal, spends zero, and never falls back to Claude. (must_have)
• Carol's WhatsApp lane and the intake/strategy lane still resolve to OpenAI and are unaffected by the disconnect. (must_have)
• Reconnecting the pipeline to a working model is a single switch command. (must_have)
• A regression test proves a disconnected fleet refuses with no Claude leak while Carol's lane stays on OpenAI. (must_have)