# Crypto_Agent Phase 6/7 v1 closure pattern

Use this reference when continuing the user's 260629 Crypto_Agent roadmap after Market Snapshot, Confluence, Entry Improvement, and Paper Lifecycle v1 are already in place.

## Phase 6 — strategy sleeves v1

Goal: separate evidence and readiness by sleeve so intraday v76 results cannot be mistaken for swing or portfolio readiness.

Pattern:

1. Add a small pure/read-only sleeve registry module, e.g. `src/strategies/strategy_sleeves.py`.
2. Define explicit sleeves:
   - `intraday_v76_strict` → existing v76 strict paper-only candidate.
   - `swing_trend_retest` → `research_only` placeholder until its own paper journal exists.
   - `portfolio_holding_advisor` → `advisor_only` placeholder until portfolio context exists.
3. Build `strategy_sleeves.v1` scorecard from true lifecycle scorecard for intraday only.
4. Mark swing and portfolio as `Research`/`Advisor` with blockers such as `swing_strategy_not_implemented`, `no_sleeve_specific_scorecard_yet`, `advisor_only_no_execution`, `portfolio_context_not_connected`.
5. Hard-code/report `live_allowed=false`, `live_order_allowed=false`, `mainnet_signed_action=false` for every sleeve.
6. Add the formatted sleeve line to the daily report and update README/phase-status docs.
7. Verify with targeted tests + full suite + daily-report smoke.

Core pitfall: do not reuse v76 intraday lifecycle evidence to promote swing or portfolio sleeves. The phase closes only as an evidence-separation milestone, not live readiness.

## Phase 7 — copytrading research v1

Goal: close copytrading as read-only research infrastructure, not live-copy readiness.

Pattern:

1. Keep runtime data outside Git under `~/.local/state/CryptoTradingBot/copy_research` or the configured `CopyResearchPaths` root.
2. Aggregate existing ctb_copy components into a concise phase scorecard:
   - runtime watchlist for wallets/vaults,
   - vault/wallet snapshot collectors,
   - discovery registry to avoid survivorship bias,
   - leader scorecard,
   - follower fill model,
   - read-only shadow-copy position engine,
   - shadow portfolio/journals,
   - copy daily report.
3. Add a `copytrading_research.v1` report module that reads runtime stats and reports evidence counts: enabled wallets/vaults, registry total, shadow decisions, allowed/blocked counts, top block reasons, data quality.
4. Always force safety fields:
   - `research_only=true`
   - `copy_permission=false`
   - `live_copy_allowed=false`
   - `vault_deposits_allowed=false`
   - `wallet_execution_allowed=false`
   - `mainnet_signed_action=false`
5. Add blockers for missing evidence, e.g. `no_enabled_copy_watchlist_entries`, `no_shadow_position_deltas_yet`, `no_leaders_in_discovery_registry_yet`, and alert on unexpected `copy_permission=true`.
6. Add the Copytrading Research line to the main HyperLiquid daily report, update README and phase-status docs.
7. Verify with copy-specific tests, full suite, copy daily-report smoke, main report smoke, unsafe-pattern scan, token-safe commit/push.

Core pitfall: `allowed` in a shadow-copy decision means only a hypothetical shadow gate passed; it is not permission to copy live, deposit into vaults, sign, or execute through any wallet/API wallet.