fix(gateway): unblock task claim; full Phase 0/1/2 remediation

Resolves the 100% claim-failure rate introduced by the gateway rewrite
  (commit 62bda0c plus 78 follow-ups). Live smoke runs hit
  `404 /api/v2/flow/developer/...` on every dev verb plus a manifest
  fallback that silently exposed off-role verbs to PMs — confirmed
  firing simultaneously in NAS agent logs (be-dev-1, be-pm, main-pm).

  Audit reports under docs/internal/audit_2026_05_04/ catalogue 49
  defects across gateway, services, prompts, MCP transport, substrate,
  and tests (8 detail reports + master synthesis). Six smoking guns;
  three proven in production logs.

  Phase 0 — unblock claim:
  - URL prefix /api/v2/flow/dev → /developer; slug-map board roles
    (product_owner, head_marketing) → /board (D-01)
  - _i_will_work_on AttributeError on None across pending /
    needs_revision / claimed re-entry branches (D-02)
  - Seed last_heartbeat_at in _qa_or_doc_claim (D-03)
  - Drop misleading i_have_committed verb; dev flow uses commit() (D-04)
  - Manifest mount via compose; flow_server + do_server fail loud
    instead of exposing all-verbs fallback (D-12)
  - MCP _post() surfaces envelope body on 4xx so agents see remediate
    hints (D-13)
    on git failure so retries aren't blocked by half-state (S-01)

  Phase 1 — lifecycle stability:
  - _resolve_skill falls back to AgentTable.capabilities (D-06)
  - main_pm_complete uses kwargs for escalate_to_ceo (D-07)
  - i_am_done auto-runs submit_verification when in_progress (D-08)
  - active_claimant_id wired in claim/unclaim paths — single-claimant
    invariant now functional (D-05)
  - qa_pass/qa_fail assert claimed_by parity with qa_agent_id (D-18)
  - Prompt-drift sweep: fail() shape, i_am_done(task_id, notes),
    subtask cap (12 hard / 8 soft), error-code symbology rewritten in
    base.md + per-role anti-patterns (D-10/11/29/30/31, D-37)

  Phase 2 — invariants + architecture:
  - Real-DB integration test exercising claim → in_progress → commit
    → submit_for_qa → i_am_done → awaiting_qa (P2-1)
  - choreographer.py → package; 3 of 6 role mixins extracted
    (board, doc, qa). _impl.py 2,526 → 2,080 lines (-18%). Continuation
    plan in docs/internal/audit_2026_05_04/p2_2_decompose_plan.md (P2-2)
  - Closure guards consolidated via _subtasks_not_terminal_envelope (P2-3)
  - TaskService.unclaim_for_reaper routed through canonical
    _validate_and_set_status; in_progress → pending added to
    VALID_TRANSITIONS (P2-4)
  - Dead code removed: i_am_done_with_catchup verb, _run_catch_up helper
    (P2-5)
  - 6 state-machine invariants asserted via property test (P2-6)
  - attempt_id (uuid4) stamped on every gateway.rejected audit row (P2-7)
  - _reconcile_orphan_claims_on_startup rolls back tasks left CLAIMED
    with branch_name=NULL from prior crashes (P2-8)
  - scripts/regenerate_verb_tables.py introspects Pydantic schemas +
    role_config; compose_prompt injects per-role tables as a layer.
    Eliminates the prompt-drift class structurally (P2-9)

  Other:
  - D-48: orchestrator mounts host's ~/.claude.json when present so
    agents don't boot from backup recovery on every spawn
  - D-49: dev dispatcher rejects role-mismatched spawns (e.g. doc task
    assigned to dev agent)

  Tests: 553 pass · ruff + mypy clean. Live NAS smoke verification
  pending — needs the stack brought back up.

agents/prompts/roles/qa.md

@@ -36,7 +36,7 @@ A pass without evidence is a betrayal of your role: the entire downstream chain
 5. For each acceptance criterion: confirm there is a referencing artifact (commit, progress entry, or file change) AND that the change actually meets it.
 6. Run tests/lint via `Bash` if your role permits; otherwise rely on the diff.
 7. `note(scope='learning', text="<what worked / what would have caught the issue earlier>")`.
-8. Pass: `pass(task_id, notes="<>=80 chars: what you reviewed, what you confirmed, any caveats>")`. Fail: `fail(task_id, issues=[{criterion, file, line, expected, actual}, ...])`.
+8. Pass: `pass(task_id, notes="<>=80 chars: what you reviewed, what you confirmed, any caveats>")`. Fail: `fail(task_id, issues=["<concrete actionable issue>", "<another>", ...])` — each issue is a single string. Reference criterion id + file + line + expected vs actual inside the string itself.
 
 ## Anti-patterns
 
@@ -44,8 +44,8 @@ A pass without evidence is a betrayal of your role: the entire downstream chain
 - ❌ Approving without reading the diff. The gateway tracks whether you called `claim_review` / `evidence`; it can detect a `pass` without evidence inspection. Fix: always re-read the diff before passing, even if the task looks trivial.
 - ❌ Running `Bash git diff` or `Bash gh pr view` to inspect changes. The PR data is already in `claim_review`'s response, and direct git/curl is denied. Call `evidence(task_id)` if you need more.
 - ❌ Trying to fix the issue yourself by editing files. You have no `Edit`/`Write` for non-trivial fixes; if you find a bug, fail with the issue list and let the developer fix it.
-- ❌ Reviewing your own work. The gateway rejects with `SELF_REVIEW_FORBIDDEN` if you were the original developer. If this happens, escalate so a different QA picks it up.
-- ❌ Passing with `notes` < 80 chars. Server-side gate rejects with `QA_NOTES_REQUIRED`.
+- ❌ Reviewing your own work. If you were the original developer, escalate so a different QA picks it up. (Self-review enforcement is best-effort at the gateway today; the convention still holds.)
+- ❌ Passing with `notes` < 80 chars. The gateway returns a `tracing_gap` envelope with `missing` containing `qa_notes>=min`.
 - ❌ Skipping the `journal:learning` entry. The gateway will reject `pass`/`fail` with a tracing-gap envelope until you've recorded one.
 
 ## When the gateway returns an error