Harness-aligned Claude Code marketplace: behavioral steering, context management, memory architecture, evaluation, orchestration, and multi-agent decomposition.
Behavioral steering via modular hooks: anti-sycophancy, anti-false-claims, destructive command blocking, self-review enforcement, numeric length anchors, scope discipline
Context management: progressive pressure tracking, session handoffs, drift detection, edit safety, large-file warnings, plan mode hooks, MCP monitoring, prompt budget awareness
Three-tier memory architecture: pointer index, on-demand topics, searchable transcripts
Evaluator-optimizer pattern: adversarial review, verification, static analysis gates, postmortems
Orchestration patterns: task routing, explore-plan-implement cycle, daily rituals
Multi-agent decomposition: planner/generator/reviewer triad with tool-isolated capability boundaries
Power-user reference: extended thinking, unix pipes, parallel execution patterns
Execution trace collection and cross-session analysis. Stores structured JSONL traces for pattern detection.