← Back to Updates ·
Comparison page →
Plan: Real-Money Asset-Class Edge Ramp
Current review first
- Freshness risk is real right now. The local canonical payload in
dashboard_data.json is timestamped 2026-05-10T04:03Z, so any "latest" verdict must be treated as stale until the next successful refresh cycle.
- The one-shot audit in
money_maker_ready_2026-05-10T05Z.md shows the core blockers are still drift alert, dragger systems, and missing walk-forward coverage in some classes.
- Data lineage confirmed:
.github/workflows/audit-dashboard.yml runs the generator, commits/deploys refreshed payload files (including dashboard_data.json), and syncs portfolio data to MySQL.
- Important DB-truth nuance: live
/audit read path remains the stocks DB, while backtests DB is secondary/transitionary per audit_trail/mysql_client.py.
Skill promotion question — answer
- No new promotion needed. A live skill already exists at
.claude/skills/money-maker-ready/SKILL.md; the proposal artifact still exists at docs/proposals/money-maker-ready_skill_proposal_2026-05-10.md.
- Recommended action: consolidation — mark proposal as superseded, keep live skill as single source of truth.
Phase 0 (Day 0-2) — Refresh-and-lock source of truth
- Gate all decisions on fresh payload age (≤2h), then rerun money-maker-ready baseline report from fresh data.
Phase 1 (Week 1) — Immediate P0 controls after approval
- Quarantine confirmed draggers.
- Mark dead systems inactive.
- Enforce drift auto-pause gate.
- Reconcile dashboard threshold display with charter thresholds in
docs/PERFORMANCE_CHARTER.md.
Phase 2 (Day 7-30) — Per-asset-class measurable lift
- CRYPTO: reduce low-PF source share and concentration risk; keep only Tier-2-eligible curated sleeve for capital simulation.
- FOREX: mutation-before-kill rehab path first (no silent kill); isolate resolver/noise vs true edge.
- COMMODITY: preserve PF edge while lifting WR and validating walk-forward coverage.
- EQUITY / ETF / BOND: push sample size to n≥100 floor where missing; prioritize ETF expansion if OOS leadership persists.
Phase 3 (Day 30-60) — Readiness conversion
- Add class-level Go/No-Go scoreboard to audit outputs with explicit gates: fresh, drift, tier, n, walkforward, concentration.
- Promote only classes that pass all hard gates for 2 consecutive weekly snapshots.
Phase 4 (Day 60-90) — Scale and hardening
- Increase capital only for classes holding Tier-2 + drift stability.
- Run weekly reruns and monthly deep-dives for regressions.
Relevant files
audit_dashboard/data/dashboard_data.json.github/workflows/audit-dashboard.ymlaudit_trail/dashboard_generator.pyaudit_trail/mysql_client.py.claude/skills/money-maker-ready/SKILL.mddocs/proposals/money-maker-ready_skill_proposal_2026-05-10.mdreports/money_maker_ready_2026-05-10T05Z.mddocs/PERFORMANCE_CHARTER.mddocs/MUTATION_THREE_AXIS_PROTOCOL.mddocs/STRATEGY_INVESTIGATION_BEFORE_KILL.md
Verification
- Confirm fresh payload and walkforward timestamps before any verdict.
- Produce a new money-maker-ready report with all sections populated.
- Validate class tiering against charter PF/WR/MDD/n thresholds.
- Compare two consecutive weekly snapshots; only passing classes become pilot-eligible.
- Verify workflow run health and that refreshed payload is deployed to live audit endpoint.
Decisions locked
- Conservative pilot gate selected: require 2 consecutive weekly Tier-2 passes.
- Rollout order selected: ETF + COMMODITY first; CRYPTO curated sleeve in parallel.
Plan persisted at /memories/session/plan.md and ready for handoff to implementation.