@hyperb1iss

implement covers verification-driven coding with tight feedback loops, distilled from 21k+ tracked operations and 600+ debugging sessions. emphasizes testing every 2-3 edits to catch quality gaps early.

decomposition framework emphasizing verification-driven task breakdown and persistent tracking across context windows. derives patterns from 200+ planning sessions to surface concrete, checkable steps over abstract bullets.

orchestrate provides pattern taxonomy for multi-agent dispatch across research swarms, parallel builds, and sequential pipelines, grounded in 597+ production codebases but lacking concrete procedural steps.

codex-review invokes cross-model validation where codex audits claude-generated code. avoids self-review bias by leveraging different training distributions and architectural assumptions.

brainstorm applies the double diamond model to ideation, separating divergent (ai-driven volume) from convergent (human judgment) phases, with persistent memory to avoid redundant exploration.

research structures multi-agent knowledge gathering in waves before synthesis, with calibration modes (quick, standard, deep) tuned to question scope. lacks concrete procedures and trigger clarity.

This skill should be used when designing terminal user interfaces, creating TUI layouts, choosing TUI color schemes, implementing keyboard navigation, building…