Agent Claude Code
Makes project or epic architecture decisions. Use inside architecture or skeptical-review phases for frontier trade-off judgment.
Agent Claude Code
Makes project or epic architecture decisions. Use inside architecture or skeptical-review phases for frontier trade-off judgment.
Agent Claude Code
Attacks planning or implementation independently. Use during analyze or speck-audit after authorship and before downstream work.
Agent Claude Code
Implements planned story work in isolation. Use during story-implement or harden after executable tasks and prerequisites are ready.
Agent Claude Code
Locates files, symbols, and boundaries quickly. Use during discovery, scan, graph repair, or recheck before deeper code analysis.
Agent Claude Code
Researches current external facts. Use inside just-in-time research or frontier scans before an external claim or choice locks.
Agent Claude Code
Extracts code behavior and reusable patterns. Use during brownfield scan, discovery, graph repair, or recheck after file boundaries are known.
Agent Claude Code
Authors project, epic, or story specifications. Use inside specify or epic-breakdown phases before planning begins.
Agent Claude Code
Judges runtime outcomes and readiness. Use at story, epic, or project prove gates after the required adversarial audit.