Agent
The Detective - scientific debugging, hypothesis testing, evidence-based root-cause analysis.
A spec-driven development platform for OpenCode that guides teams through a structured 5‑phase workflow (Discuss → Plan → Execute → Audit → Confirm), with built‑in delegation, checkpoints, and verification.
Agent
The Detective - scientific debugging, hypothesis testing, evidence-based root-cause analysis.
Agent
Frontend high-tier executor for design-sensitive UI work — component architecture, UX, accessibility, visual polish.
Agent
Frontend low-tier executor for UI mechanical tasks — markup, simple styling, copy.
Agent
Frontend medium-tier executor for standard component work, UI logic/state wiring, and moderate refactors within existing patterns.
Agent
High-tier executor for architecture, complex algorithms, API design, and security-sensitive work.
Agent
Low-tier executor for mechanical tasks — config edits, renaming, scaffolding, markdown.
Agent
Medium-tier executor for business logic, utilities, tests, refactoring, and scripting.
Agent
The Scout - fast codebase mapping, pattern detection, terrain reconnaissance.
Agent
The Conductor. Coordinates all work, NEVER writes code, maintains clean context, enforces gates.
Agent
The Architect - creates SPEC/BLUEPRINT with wave decomposition, traceability, and verification criteria.
Agent
The Scholar - deep domain research, technology evaluation, synthesis.
Agent
The Guardian - test writing, QA, coverage thinking, edge cases.
Agent
The Auditor - ruthless verification against spec, security focus, trust nothing.
Agent
The Wave Auditor - verifies exactly one wave's deliverables against its tasks, records gaps as verification rows.
Agent
The Scribe - documentation generation, technical writing, clarity.