Free Will Pivot

Do NOT run the entire test suite as a confirmation step — that wastes the block.

Step 2: Discovery Cascade

Execute this priority-ordered cascade. Stop at the first category that yields actionable work:

Priority 1: Queued Tasks

ls TASKS/*.md  # pending tasks (not .done, .failed, .inprogress)

If any pending tasks exist, execute the highest-priority one using the task-execution skill.

Priority 2: System Health Issues

cat HEARTBEAT.md  # check for any non-healthy items
cat STATE/progress_history.json | tail -5  # recent scores and trends

Look for:

• Any heartbeat check showing degraded/failing status
• Progress dimensions marked degrading or below 70
• Metric anomalies (failure rates, stale data, unbounded growth)
• Service issues (high memory, stale PIDs, log bloat)

Priority 3: Overdue Maintenance

Scan for systemic hygiene issues:

• STATE/ file sizes — anything over 5MB needs rotation or pruning
• Log retention — run scripts/log_retention.py if logs are large
• Test pollution — check if test runs are contaminating production state files
• Dead code / unused files — stale .inprogress files, orphaned configs
• Dependency health — outdated packages, security advisories

Priority 4: Code Quality

• Run ruff or linting on recently changed files
• Fix any test failures in the suite (run targeted, not full suite)
• Address TODO/FIXME/HACK comments in actively maintained files

Priority 5: NovaTrade Research & Advancement

If the system is healthy and no maintenance is needed, invest in NovaTrade:

• Review recent trade journal entries for patterns
• Research strategy improvements based on rejection reasons
• Investigate execution pipeline gaps
• Advance any open NovaTrade implementation phases

Priority 6: Infrastructure Improvement

If everything above is clean:

• Improve monitoring coverage
• Add missing test cases for edge cases discovered in prior sessions
• Enhance error messages or logging in problem areas
• Document architectural decisions that lack ADRs

Step 3: Execute the Pivot Work

Once a target is identified:

1. State your pivot reason — one sentence explaining why you chose this work over alternatives.
2. Execute with full rigor — treat the pivot work with the same discipline as an assigned task. Use implementation-team patterns for non-trivial changes.
3. Write tests if modifying production code.
4. Write output to OUTPUT/<original_stem>__<timestamp>.md preserving the shift block lineage.

Step 4: Output Contract

Every pivot execution must end with:

## CONTRACT
pivot_reason: <why this work was chosen over alternatives>
discovery_source: <which cascade priority yielded the work — tasks/health/maintenance/quality/novatrade/infra>
work_performed: <one-line description>